Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

32
IEEE COMMUNICATIONS SURVEYS & TUTORIALS, ACCEPTED FOR PUBLICATION 1 Cloud-Based Augmentation for Mobile Devices: Motivation, Taxonomies, and Open Challenges Saeid Abolfazli, Member, IEEE, Zohreh Sanaei, Member, IEEE, Ejaz Ahmed, Member, IEEE, Abdullah Gani, Senior Member, IEEE, Rajkumar Buyya, Senior Member, IEEE Abstract—Recently, Cloud-based Mobile Augmentation (CMA) approaches have gained remarkable ground from academia and industry. CMA is the state-of-the-art mobile augmentation model that employs resource-rich clouds to increase, enhance, and optimize computing capabilities of mobile devices aiming at execution of resource-intensive mobile applications. Augmented mobile devices envision to perform extensive computations and to store big data beyond their intrinsic capabilities with least footprint and vulnerability. Researchers utilize varied cloud- based computing resources (e.g., distant clouds and nearby mobile nodes) to meet various computing requirements of mobile users. However, employing cloud-based computing resources is not a straightforward panacea. Comprehending critical factors (e.g., current state of mobile client and remote resources) that impact on augmentation process and optimum selection of cloud- based resource types are some challenges that hinder CMA adaptability. This paper comprehensively surveys the mobile aug- mentation domain and presents taxonomy of CMA approaches. The objectives of this study is to highlight the effects of remote resources on the quality and reliability of augmentation processes and discuss the challenges and opportunities of employing varied cloud-based resources in augmenting mobile devices. We present augmentation definition, motivation, and taxonomy of augmen- tation types, including traditional and cloud-based. We critically analyze the state-of-the-art CMA approaches and classify them into four groups of distant fixed, proximate fixed, proximate mobile, and hybrid to present a taxonomy. Vital decision making and performance limitation factors that influence on the adoption of CMA approaches are introduced and an exemplary decision making flowchart for future CMA approaches are presented. Im- pacts of CMA approaches on mobile computing is discussed and open challenges are presented as the future research directions. Index Terms—Cloud-based Mobile Augmentation, Mobile Cloud Computing, Cloud Computing, Resource-intensive Mobile Application, Computation Offloading, Resource Outsourcing. I. I NTRODUCTION S INCE a decade ago, the visions of ‘information under fingertip’ [1] and ‘unrestricted mobile computing’ [2] aroused the need to enhance computing power of mobile devices to meet the insatiable computing demands of mobile users [3]. In the late 90s, the concept of load sharing and Manuscript received Dec 18, 2012; revised March 05, 2013 and 06 May, 2013;This work is funded by the Malaysian Ministry of Higher Education under the University of Malaya High Impact Research Grant - UM.C/HIR/MOHE/FCSIT/03. Ejaz Ahmed’s research work is supported by the Bright Spark Unit, University of Malaya, Malaysia. Saeid Abolfazli(corresponding author), Zohreh Sanaei, Ejaz Ahmed, and Abdullah Gani are with the Department of Computer System & Technology, The University of Malaya, Kuala Lumpur, Malaysia (e-mail: {abolfazli,sanaei, ejazahmed}@ieee.org; [email protected]) RajKumar Buyya is with the Department of Computing and Information Systems, The University of Melbourne, 111, Barry Street, Carlton, Melbourne, VIC 3053, Australia, Email: [email protected] remote execution aimed to augment computing capabilities of mobile devices by shifting the resource-intensive mobile codes to surrogates (powerful computing device(s) in vicinity) [4]– [6]. Although remote execution efforts [7]–[18] have yielded many impressive achievements, several challenges such as reliability, security, and elasticity of surrogates hinder the remote execution adaptability [19]. For instance, the resource sharing and computing services of surrogates can be termi- nated without prior notice and their content can be accessed and altered by the surrogate machine or other users in the absence of a Service Level Agreement (SLA). SLA is a formal contract employed and negotiated in advance between service provider and consumer to enforce certain level of quality against a fee. Few years later, emergence of cloud resources created an opportunity to mitigate the shortcomings of utilizing surro- gates in augmenting mobile devices. Cloud is a type of dis- tributed system comprised of a cluster of powerful computers accessible as unified computing resource(s) based on an SLA [20]. Cloud computing as “a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of con- figurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service or service provider interaction” [21] stimulates researchers to adopt the cutting edge technology in mobile device augmenta- tion: Cloud-based Mobile Augmentation (CMA). Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud computing technolo- gies and principles to increase, enhance, and optimize com- puting capabilities of mobile devices by executing resource- intensive mobile application components in the resource-rich cloud-based resources. Cloud-based resources include varied types of mobile/immobile computing devices that follow cloud computing principles [22], [23] to perform computations on behalf of the resource-constraint mobile devices. Figure 1 depicts major building blocks of a typical CMA system. It is notable that these building blocks are optional superset, and specific CMA system may not have all these building blocks. CMA efforts [24]–[27], [27]–[49] exploit various cloud- based computing resources, especially distant clouds and prox- imate mobile nodes to augment mobile devices. Distant clouds are giant clouds such as Amazon EC2 1 located inside the vendor premise —far away from the mobile clients—offering infinite, elastic computing resources with extreme computing 1 http://aws.amazon.com/ec2/

description

Comprehensive Survey on Mobile Cloud Computing. The paper abstract is here: Recently, Cloud-based Mobile Augmentation (CMA) approaches have gained remarkable ground from academia and industry. CMA is the state-of-the-art mobile augmentation model that employs resource-rich clouds to increase, enhance, and optimize computing capabilities of mobile devices aiming at execution of resource-intensive mobile applications. Augmented mobile devices envision to perform extensive computations and to store big data beyond their intrinsic capabilities with least footprint and vulnerability. Researchers utilize varied cloud-based computing resources (e.g., distant clouds and nearby mobile nodes) to meet various computing requirements of mobile users. However, employing cloud-based computing resources is not a straightforward panacea. Comprehending critical factors (e.g., current state of mobile client and remote resources) that impact on augmentation process and optimum selection of cloud-based resource types are some challenges that hinder CMA adaptability. This paper comprehensively surveys the mobile augmentation domain and presents taxonomy of CMA approaches. The objectives of this study is to highlight the effects of remote resources on the quality and reliability of augmentation processes and discuss the challenges and opportunities of employing varied cloud-based resources in augmenting mobile devices. We present augmentation definition, motivation, and taxonomy of augmentation types, including traditional and cloud-based. We critically analyze the state-of-the-art CMA approaches and classify them into four groups of distant fixed, proximate fixed, proximate mobile, and hybrid to present a taxonomy. Vital decision making and performance limitation factors that influence on the adoption of CMA approaches are introduced and an exemplary decision making flowchart for future CMA approaches are presented. Impacts of CMA approaches on mobile computing is discussed and open challenges are presented as the future research directions.

Transcript of Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

Page 1: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 1

Cloud-Based Augmentation for Mobile DevicesMotivation Taxonomies and Open Challenges

Saeid AbolfazliMember IEEEZohreh SanaeiMember IEEEEjaz AhmedMember IEEEAbdullahGani Senior Member IEEE Rajkumar BuyyaSenior Member IEEE

AbstractmdashRecently Cloud-based Mobile Augmentation (CMA)approaches have gained remarkable ground from academia andindustry CMA is the state-of-the-art mobile augmentationmodelthat employs resource-rich clouds to increase enhance andoptimize computing capabilities of mobile devices aiming atexecution of resource-intensive mobile applications Augmentedmobile devices envision to perform extensive computationsandto store big data beyond their intrinsic capabilities with leastfootprint and vulnerability Researchers utilize varied cloud-based computing resources (eg distant clouds and nearbymobile nodes) to meet various computing requirements of mobileusers However employing cloud-based computing resources isnot a straightforward panacea Comprehending critical factors(eg current state of mobile client and remote resources)thatimpact on augmentation process and optimum selection of cloud-based resource types are some challenges that hinder CMAadaptability This paper comprehensively surveys the mobile aug-mentation domain and presents taxonomy of CMA approachesThe objectives of this study is to highlight the effects of remoteresources on the quality and reliability of augmentation processesand discuss the challenges and opportunities of employing variedcloud-based resources in augmenting mobile devices We presentaugmentation definition motivation and taxonomy of augmen-tation types including traditional and cloud-based We criticallyanalyze the state-of-the-art CMA approaches and classify theminto four groups of distant fixed proximate fixed proximatemobile and hybrid to present a taxonomy Vital decision makingand performance limitation factors that influence on the adoptionof CMA approaches are introduced and an exemplary decisionmaking flowchart for future CMA approaches are presented Im-pacts of CMA approaches on mobile computing is discussed andopen challenges are presented as the future research directions

Index TermsmdashCloud-based Mobile Augmentation MobileCloud Computing Cloud Computing Resource-intensive MobileApplication Computation Offloading Resource Outsourcing

I I NTRODUCTION

SINCE a decade ago the visions oflsquoinformation underfingertiprsquo [1] and lsquounrestricted mobile computingrsquo[2]

aroused the need to enhance computing power of mobiledevices to meet the insatiable computing demands of mobileusers [3] In the late 90s the concept of load sharing and

Manuscript received Dec 18 2012 revised March 05 2013 and06May 2013This work is funded by the Malaysian Ministry of HigherEducation under the University of Malaya High Impact Research Grant -UMCHIRMOHEFCSIT03 Ejaz Ahmedrsquos research work is supported bythe Bright Spark Unit University of Malaya Malaysia

Saeid Abolfazli(corresponding author) Zohreh Sanaei Ejaz Ahmed andAbdullah Gani are with the Department of Computer System amp TechnologyThe University of Malaya Kuala Lumpur Malaysia (e-mailabolfazlisanaeiejazahmedieeeorg abdullahumedumy)

RajKumar Buyya is with the Department of Computing and InformationSystems The University of Melbourne 111 Barry Street Carlton MelbourneVIC 3053 Australia Email rajcsseunimelbeduau

remote execution aimed to augment computing capabilities ofmobile devices by shifting the resource-intensive mobile codesto surrogates (powerful computing device(s) in vicinity) [4]ndash[6] Although remote execution efforts [7]ndash[18] have yieldedmany impressive achievements several challenges such asreliability security and elasticity of surrogates hinder theremote execution adaptability [19] For instance the resourcesharing and computing services of surrogates can be termi-nated without prior notice and their content can be accessedand altered by the surrogate machine or other users in theabsence of a Service Level Agreement (SLA) SLA is a formalcontract employed and negotiated in advance between serviceprovider and consumer to enforce certain level of qualityagainst a fee

Few years later emergence of cloud resources created anopportunity to mitigate the shortcomings of utilizing surro-gates in augmenting mobile devices Cloud is a type of dis-tributed system comprised of a cluster of powerful computersaccessible as unified computing resource(s) based on an SLA[20] Cloud computing as ldquoa model for enabling ubiquitousconvenient on-demand network access to a shared pool of con-figurable computing resources (eg networks servers storageapplications and services) that can be rapidly provisionedand released with minimal management effort or service orservice provider interactionrdquo [21] stimulates researchers toadopt the cutting edge technology in mobile device augmenta-tion Cloud-based Mobile Augmentation (CMA) Cloud-basedMobile Augmentation (CMA) is the-state-of-the-art mobileaugmentation model that leverages cloud computing technolo-gies and principles to increase enhance and optimize com-puting capabilities of mobile devices by executing resource-intensive mobile application components in the resource-richcloud-based resources Cloud-based resources include variedtypes of mobileimmobile computing devices that follow cloudcomputing principles [22] [23] to perform computations onbehalf of the resource-constraint mobile devices Figure 1depicts major building blocks of a typical CMA system Itis notable that these building blocks are optional superset andspecific CMA system may not have all these building blocks

CMA efforts [24]ndash[27] [27]ndash[49] exploit various cloud-based computing resources especially distant clouds and prox-imate mobile nodes to augment mobile devices Distant cloudsare giant clouds such as Amazon EC21 located inside thevendor premise mdashfar away from the mobile clientsmdashofferinginfinite elastic computing resources with extreme computing

1httpawsamazoncomec2

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 2

Fig 1 Major Building Blocks of an Exemplary CMA System

power and high WAN (Wide Area Network) latency Proximatemobile nodes are building a cluster of mobile computing de-vices scattered near the mobile clients offer limited computingpower with lower WAN latency than distant clouds

Although heterogeneity among cloud-based resources in-creases service flexibility and enhances usersrsquo computing expe-rience determining the most appropriate computing resourcesamong available options and performing upfront analysis ofinfluential factors (eg user preferences and available nativemobile resources) are critical in the adaptability of CMAapproaches Thus lsquoresource schedulerrsquo and lsquoanalyzer andoptimizerrsquo components depicted in Figure 1 are needed toanalyze and allocate appropriate resources to each task ina typical CMA system Moreover several questions need tobe addressed before the CMA concept can be successfullyemployed in the real scenarios For instance can CMA aug-ment computing capabilities of mobile devices and save localresources to enhance user experience Is CMA always feasibleand beneficial Which type of resources is appropriate for acertain task Answering these questions requires lsquomonitoringand profilerrsquo lsquoQoS managementrsquo lsquocontext managementrsquo andlsquodecision making enginersquo components to perform in eachCMA system (see Figure 1) Therefore an augmentation deci-sion engine similar to those used in [25] [33] [49] and exem-plary decision making flow presented in this paper (discussedin part VI-C) to determine the mobile augmentation feasibilityis needed to amend the CMA performance and reliabilityDuring augmentation process the local and native applicationstate stack needs synchronization to ensure integrity betweennative and remote data Upon successful outsourcing remoteresults need to be returned and integrated to the source mobiledevice Thus the lsquoSynchronizerrsquo component needs to performin typical CMA approaches (see Figure 1)

Although CMA approaches can empower mobile process-ing and storage capabilities several disadvantages such asapplication development complexity and unauthorized accessto remote data demand a systematized plenary solutionPerformance of the CMA systems is highly influenced byvarious challenges and issues of wireless networking andcloud computing technologies CMA researchers require a

high performance elastic robust reliable and foreseeablecommunication throughput between mobile nodes and cloudservers which is not yet realized despite of remarkable effortsand achievements of communication and networking societiesCurrent shortcomings and deficiencies of wireless communi-cation and networking especially seamless connectivity andmobility high performance communication throughput pro-visioning and wireless data interception discourage systemanalysts engineers developers and entrepreneurs from de-ploying CMA-enabled mobile applications due to the high riskof system malfunction and user experience degradation

Moreover CMA systems require accurate estimation mech-anisms to predict the overall time and energy consumptionof communication and computation tasks while exploitingclouds Such estimation is a challenging task consideringhuge infrastructuresrsquo performance diversity [50] and policyheterogeneity [51] of cloud services in intermittent wirelessenvironment Despite of blooming efforts endeavoring to ana-lyze and comprehend the cloud computing model and behavior[52]ndash[55] CMA solutions are still unable to accurately fore-see required time and energy of exploiting cloud resourcesto execute intensive applications Additionally sundry cloudchallenges especially live VM migration infrastructureandplatform heterogeneity efficient allocation of clouds to tasksQoS management security privacy and trust in cloud increasesystem complexity and decrease successful CMA systemsadoption

Among limited studies of the domain [19] and [56] surveyremote execution and application offloading algorithms withfocus on how task offloading is performed in various effortsFernando et al [57] and Dinh et al [58] sought to explain theconvergence of mobile and cloud computing and distinguishit from the earlier domains such as cloud and grid computing[59] The authors describe issues particularly mobile applica-tion offloading privacy and security context awareness anddata management Sanaei Abolfazli Gani and Buyya [51]present a comprehensive survey on MCC with major focuson heterogeneity The authors describe the challenges andopportunities imposed by heterogeneity and identify hardwareplatform feature API and network as the roots of MCC

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 3

TABLE IL IST OF ACRONYMS AND CORRESPONDINGFULL FORMS

Acronym Full form

2D 2 Dimensional

2G 2nd Generation

3D 3 Dimensional

3G 3rd Generation

API Application Programming Interface

App Application (mobile application)

ARM Advanced RISC Machines

CMA Cloud-based Mobile Augmentation

CPU Central Processing Unit

DSL Domain Specific Language

DVMS Dynamic VM Synthesis

FTP File Transfer Protocol

GPU Graphics Processing Unit

GUI Graphical User Interface

IO InputOutput

IaaS Infrastructure as a Service

IP Internet Protocol

IP TV Internet Protocol Television

iSCSI Internet Small Computer System Interface

MCC Mobile Cloud Computing

MNO Mobile Network Operator

OS Operating System

P2P Peer-to-Peer

PC Personal Computer

QoS Quality of Service

RampD Research and Development

RAM Random Access Memory

RISC Reduced Instruction Set Computing

RPC Remote Procedure Call

SAL Service Abstraction Layer

SLA Service Level Agreement

TCP Transmission Control Protocol

UDDI Universal Description Discovery and Integration

UI User Interface

VM Virtual Machine

WAN Wide Area Network

Wi-Fi Wireless Fidelity

heterogeneity They explain major heterogeneity handlingap-proaches particularly virtualization service orientedarchi-tecture and semantic technology However the computingperformance distance elasticity availability reliability andmulti-tenancy of remote resources are marginally discussedin these studies that necessitate further research to explainthe impact of remote resources on augmentation process andhighlight paradigm shift from the unreliable surrogates to

reliable clouds

In this paper we survey the state-of-the-art mobile augmen-tation efforts that employ cloud computing infrastructures toenhance computing capabilities of resource-constraint mobiledevices especially smartphones To the best of our knowledgethis is the first effort that studies the impacts of cloud-basedcomputing resources on mobile augmentation process We dif-ferentiate augmentation from similar concepts of load sharingand remote execution and present augmentation motivationWe review efforts that endeavor to mitigate the mobile devicesrsquoshortcomings and classify them as hardware and software todevise a taxonomy The impacts of CMA in mobile comput-ing are presented The characteristics of cloud-based remoteresources and their role in CMA effectiveness are studied andclassified into four groups namely distant immobile cloudsproximate immobile computing entities proximate mobilecomputing entities and hybrid based on their mobility andphysical location traits Furthermore the state-of-the-art CMAmodels are reviewed and taxonomized into four classes ofdistant fixed proximate fixed proximate mobile and hybridaccording to our cloud-based resource classification Factorsimpact on the CMA adaptability are identified and described asaugmentation environment user preferences and requirementsmobile devices cloud servers and contents Five major metricsthat limit the performance of CMA approaches are studied Asample flowchart of decision making engines for imminentCMA solutions is presented and several open challenges arediscussed as the future research directions Such survey isbeneficial to the communication and networking communitiesbecause comprehending CMA process and current deploy-ment challenges are beneficial in modifying the fundamentalnetworking infrastructures to optimize the CMA process Inthis paper we use the terms mobile devices and smartphonesinterchangeably with similar notion Table I shows the listofacronyms used in the paper

The remainder of this paper is organized as follows SectionII introduces mobile computation augmentation presents itsmotivation and describes the taxonomy of mobile augmenta-tion types The impacts of CMA on mobile computing arepresented in Section III Section IV presents the analysisand taxonomy of varied cloud-based augmentation resourcesComprehensive survey of the state-of-the-art CMA approachesis presented and taxonomy is devised in Section V Wediscuss the CMA decision making and limitation factors andillustrate CMA feasibility in Section VI Finally open researchchallenges are presented in Section VII and paper is concludedin Section VIII

II M OBILE COMPUTATION AUGMENTATION

In this Section we present a definition on mobile computingaugmentation based on the available definitions on the relevantconcepts particularly remote execution [5] and cyber foraging[6] Additionally the motivation for performing mobile com-putation augmentation is described and taxonomy of mobileaugmentation types is presented

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 4

TABLE IIINITIAL FEATURES OFMOBILE EMPOWERMENTAPPROACHES

Approach Architecture Client Load Migration Partitioning Server Mobility

Load Sharing Client-Server Entire Task Entire task NA Server NA

RemoteClient-Server Entire Task Entirepartial Static

Server NoExecution desktop

Cyber Client-ServerEntire Task Entirepartial Dynamic

Surrogates NoForaging Peer-to-Peer

Mobile Varies eg

Nil

Entirepartial Static amp Server YesComputation Client-server Entirepartial Nil migration dynamic surrogateAugmentation P2P Adhoc (Use remote ampmobile

collaborative services)

A Definition

Indeed empowering computation capabilities of mobiledevices is not a new concept and there have been differentapproaches to achieve this goal including load sharing [4]remote execution [5] cyber foraging [6] and computationoffloading [60] [61] that are described as follows We haveanalyzed them and summarized the analysis results in TableII Results in this Table are extracted from the early efforts ineach category which are already deviated from their originalcharacteristics due to the research achievements

bull Load Sharing Othman and Hailesrsquo work [4] in 1998can be considered as one of the earliest efforts to conservenative resources of mobile devices using a software approachThe main idea is inspired from the concept of load balancingin distributed computing that is ldquoa strategy which attemptsto share loads in a distributed system without attemptingto equalize its loadrdquo [4] This approach migrates the wholecomputation job for remote execution It considers severalmetrics such as job size available bandwidth and result sizeto identify if the load balancing and transferring the job tothe remote computer can save energy However they need tosend the task and data to the nearest base station and wait forthe results to return The base station is responsible to findappropriate server to run the job and forward the results backto the mobile device Moreover computing server is a fixedcomputer and there is no provision for user and code mobilityat run time

bull Remote ExecutionThe concept of remote execution formobile clients emerged in 90s and several researchers [5][62]ndash[65] endeavor to enable mobile computers to performingremote computation and data storage to conserve their scarcenative resources and battery In 1998 [5] feasibility of remoteexecution concept on mobile computers particularly laptopswas investigated The authors report that remote executioncansave energy if the remote processing cost is lower than localexecution Remote execution involves migrating computingtasks from the mobile device to the server prior the executionThe server performs the task and sends back the results tothe mobile device However difference between computationpower of client and server is not a metric of decision makingin this method Moreover the whole task needs to be migratedto the remote server prior the execution which is an expensive

effort It also neglects the impact of environment characteris-tics on the remote execution outcome Static decision makingis another shortcoming of this proposal

bull Cyber Foraging Satyanarayana in 2001 [6] furtherdeveloped the remote execution idea by considering dynamismin remote execution process The author defined cyber forag-ing as the process ldquoto dynamically augment the computingresources of a wireless mobile computer by exploiting wiredhardware infrastructurerdquo Resources in cyber foraging arestationary computers or servers in public places mdashconnectedto wired Internet and power cablemdashthat are willing to performintensive computation on behalf of the resource-constraintmobile devices in vicinity

However load sharing remote execution and cyber forag-ing approaches assume that the whole computing task is storedin the device and hence it requires the intensive code and datato be identified and partitioned for offloading mdasheither stati-cally prior the execution or dynamically at runtime mdashwhichimpose large overhead on resource-poor mobile device [19]Moreover as Kumar et al [66] explain for each mobileuser that runs the intensive application the whole offloadingprocess must be repeated including decision making processin the device and transferring the heavy components and largedata to the network Due to slight differences among theseconcepts researchers use the terms lsquoremote executionrsquo lsquocyberforagingrsquo and lsquocomputation offloadingrsquo interchangeably in theliterature with similar principle and notion

Nevertheless researchers in recent activities [36] [42] [45][46] aim to enhance computing capabilities of mobile devicesin a slightly different manner They assume to store theintensive code and data outside the device and keep the rest inthe mobile device instead of storing the whole task mdashincludingboth lightweight and intensive code and data mdashin the mobiledevice Therefore the overhead of identifying partitioningand migrating the resource-intensive task is mitigated energyis saved and storage problem is alleviated in mobile devicesMoreover storing intensive components outside the devicein a publicly accessible storage can increase their reusabilityand enable more than one user to leverage their computationservices Therefore we coin the termmobile computationaugmentationas the wider phrase that subsumes load sharingremote execution cyber foraging and other approaches that

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 5

augment computing capabilities of mobile devicesbull Mobile Computation Augmentation Mobile computa-

tion augmentation or augmentation in brief is the processofincreasing enhancing and optimizing computing capabilitiesof mobile devices by leveraging varied feasible approacheshardware and software Mobile device is any non-stationarybattery-operating computing entity able to interact with end-user and execute transactions store data and communicatewith the environment using wireless technologies and variedsensors Smartphone Tablet handheldwearable computingdevices and vehicle mount computers are mobile device in-stances Approaches that can augment mobile devices includehardware and software Hardware approach involves manu-facturing high-end physical components particularly CPUmemory storage and battery Software approaches can bemdashbut are not limited to mdashcomputation offloading remote datastorage wireless communication resource-aware computingfidelity adaptation and remote service request (eg contextacquisition)

Augmentation approaches can increase computing capabil-ities of mobile devices and conserve energy They can beleveraged in three main categories of applications namely(i)computing-intensive software such as speech recognition andnatural language processing (ii) data-intensive programs suchas enterprise applications and (iii) communication-intensiveapplications such as online video streaming applicationsTheaugmented mobile device is able to perform complex tasks thatcould not otherwise perform Hence the mobile applicationdevelopers do not take into account resource shortcomingsof mobile devices in developing mobile application and userswill not consider their devices weaknesses in utilizing variedcomplex applications

B Motivation

Mobile devices have recently gained momentous ground inseveral communities like governmental agencies enterprisessocial service providers (eg insurance Police fire depart-ments) healthcare education and engineering organizations[67] [68] However despite of significant improvement inmobilesrsquo computing capabilities still computing requirementsof mobile users especially enterprise users is not achieved

Several intrinsic deficiencies of mobile devices encumberfeasibility of intense mobile computing and motivate aug-mentation Leveraging augmentation approaches vision ofperforming intense mobile operations and control such asremote surgery on-site engineering and visionary scenariossimilar to the lost child and disaster relief described in [69]will become reality In this part we analyze and taxonomizesmartphonesrsquo deficits that can be alleviated by augmentationFigure 2 depicts our devised taxonomy

1) Processing PowerProcessing deficiencies of mobileclients due to slow processing speed and limited RAM is oneof the major challenges in mobile computing [69] Mobile de-vices are expected to have high processing capabilities similarto computing capabilities of desktop machines for performingcomputing-intensive tasks to enrich user experience Realizingsuch vision requires powerful processor being able to performlarge number of transactions in a short course of time

Large internal memoryRAM to store state stack of allrunning applications and background services is also lackingBeside local memory limitations memory leakage also inten-sifies memory restrains of mobile devices Memory leakageis the state of memory cells being unnecessarily occupied byrunning applications and services or those cells that are notreleased after utilization Garbage-collector-based languageslike Java in Android2 contribute to memory leakage due tofailed or delayed removal of unused objects from the memory[70] Androidrsquos kernel level transactions can also leak memoryin the absence of memory management mechanisms [70][71] Moreover inward deficiency and inefficient design andimplementation of mobile applications can also waste scarcememory of mobile devices Thus in the absence of requiredmemory applications are frequently paused or terminatedby the operating system leading to longer execution timeexcessive resource dissipation and ultimately mobile userexperience degradation

2) Energy ResourcesEnergy is the only non-replenishableresource in mobile devices that demands external resourcesto be replenished [72] [73] Currently energy requirement ofa mobile device is supplied via lithium-ion battery that lastsonly few hours if device is computationally engaged Batterycapacity is increasing at about 5 to 10 a year [74] [75]as battery cells are excessively dense [72] Moreover mo-bile device manufacturers endeavor to attain device lightnesscompactness and handiness which prevent exploitation ofbulky long-lasting batteries User safety is another concern thatconfines manufacturers to produce low capacity batteries [76]While explosion of a battery with few hundreds milliamperescapacity can jeopardize human life [77] explosion of a high-capacity battery can carry catastrophic consequences

Energy harvesting efforts [78]ndash[80] seek to replenish energyfrom renewable resources particularly human movement solarenergy and wireless radiation but these resources are mostlyintermittent and not available on-demand [81] For instancea sitting mobile user at night cannot have any external powersource in the absence of wall power and wireless radiationsMoreover researchers aim at reducing the energy overheadin different aspects of computing including hardware OSapplication and networking interface [82] [83] Effortsare di-rected to develop alternative energy resources such as nuclearbatteries that will likely last months or years [84] Howeversignificant deal of RampD is needed to fulfill ever-increasingenergy requirements of mobile users

Hence in the absence of long-spanning energy on mobiledevices alternative augmentation approaches play a vitalrolein maturing mobile and ubiquitous computing

3) Local Storage Drastic increasing in the number ofapplications and amount of digital contents such as picturessongs movies and home films [85] from one hand and limitedstorage of mobile devices from the other hand decelerateusability of mobile devices While PCs are able to locallystore huge amount of data smartphones are limited to fewgigabytes of space which are mostly occupied by systemfiles user applications and personal data Therefore frequent

2urlhttpwwwandroidcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 6

Fig 2 Taxonomy of Augmentation Motivation Intrinsic andnon-intrinsic mobile challenges motivate augmentation

storing updating and deleting data as well as uninstalling andreinstalling applications due to space limitation cause irksomeimpediments to mobile users [86] Additionally deliveringoffline usability which is one of the most important character-istics of contemporary applications remains an open challengesince mobile devices lack large local storage

4) Visualization CapabilitiesEffective data visualizationon small mobile devicesrsquo screen is a non-trivial task whencurrent screen manufacturing technologies and energy limita-tions do not allow significant size extensions without losingdevice handiness Currently smartphones like HTC One X3

and Samsung Galaxy Note II4 have the biggest screens at47 and 55 inches respectively however they are very smallcompared to PCs and notebooks

Therefore efficient data visualization in small smartphonesrsquoscreen necessitates software-based techniques similar totab-ular pages 3D objects multiple desktops switching betweenlandscape and portrait views (needs accelerometer) and verbalcommunication to virtually increase presentation area Re-cently computing-intensive mobile 3D display technologyispromising to noticeably mitigate the visualization deficitofcontemporary smartphones Glass-free auto-stereoscopicdis-plays [87] can present 3D data by exploiting binocular parallaxto offer a different view for each eye Taking advantagesof current and imminent software-based techniques besidenative tools especially tilting sensors significantly improve themobile visualization capabilities in the near future Howeverthese approaches are computation-intensive processes thatquickly drain battery [87] [88] A feasible alternative solutionto realize software-based content presentation approaches is toaugment smartphonesrsquo computing capabilities

5) Security Privacy and Data SafetyMobile end-usersare concerned about security and privacy of their personaldata banking records and online behaviors [89] The dramaticincrease in cybercrime and security threats within mobiledevices [90] cloud resources [91] and wireless transactionsmakes security and privacy more challenging than ever [92]Moreover performing complex cryptographic algorithms islikely infeasible because of computing deficiencies of mobile

3httpwwwhtccomwwwsmartphoneshtc-one-x4httpwwwsamsungcommyconsumermobile-devicesgalaxy-

notegalaxy-noteGT-N7100RWDXME

devices Securing files using pair of credentials is also lessrealistic in the absence of large keyboard

Data safety is another challenge of mobile devices becauseinformation stored inside the local storage of mobile devicesare susceptible to safety breaches due to high probability ofhardware malfunction physical damage stealing and loss

Amalgam of these problems and deficiencies in mobilecomputing stimulates researchers from academia and industryto exploit novel technologies and approaches to augmentcomputing capabilities of mobile devices which is subject ofthis study

C Mobile Augmentation Types Taxonomy

In this Section we analyze and classify augmentation ap-proaches into two major types of hardware and software Ourdevised taxonomy is depicted in Figure 3 and described asfollowsHardware The hardware approach aims to empower smart-phones by exploiting powerful resources particularly multi-core CPUs with high clock speed [93] large screen and long-lasting battery [84] [94] ARM5 and Samsung6 are major mo-bile processor manufacturers producing multi-core processorssuch as ARM Cortex-597 and Samsung Exynos 5 Octa core8

that perform in higher speed than a single core processors[93] However doubling the CPU clock speed approximatelyoctuples the device energy consumption [66]

Nevertheless augmentation via sophisticated hardware ishindered by several obstacles Firstly generating powerfulprocessor large storage and big screen decrease smartphonehandiness due to additional heat size and weight Secondlyconsidering the fact that utilizing long-lasting battery in smallmobile devices is not feasible with current technologies re-source enlargement contributes toward faster battery drainageand shorter battery life time Thirdly equipping mobile de-vices with high-end hardware noticeably increases their price

5httparmcom6httpsamsungcom7httpwwwarmcomproductsprocessorscortex-a50cortex-a57-

processorphp8httpwwwsamsungcomglobalbusinesssemiconductorminisiteExynos

blog CES 2013 SamsungMobilizes Possibility with Exynos5Octahtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 2: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 2

Fig 1 Major Building Blocks of an Exemplary CMA System

power and high WAN (Wide Area Network) latency Proximatemobile nodes are building a cluster of mobile computing de-vices scattered near the mobile clients offer limited computingpower with lower WAN latency than distant clouds

Although heterogeneity among cloud-based resources in-creases service flexibility and enhances usersrsquo computing expe-rience determining the most appropriate computing resourcesamong available options and performing upfront analysis ofinfluential factors (eg user preferences and available nativemobile resources) are critical in the adaptability of CMAapproaches Thus lsquoresource schedulerrsquo and lsquoanalyzer andoptimizerrsquo components depicted in Figure 1 are needed toanalyze and allocate appropriate resources to each task ina typical CMA system Moreover several questions need tobe addressed before the CMA concept can be successfullyemployed in the real scenarios For instance can CMA aug-ment computing capabilities of mobile devices and save localresources to enhance user experience Is CMA always feasibleand beneficial Which type of resources is appropriate for acertain task Answering these questions requires lsquomonitoringand profilerrsquo lsquoQoS managementrsquo lsquocontext managementrsquo andlsquodecision making enginersquo components to perform in eachCMA system (see Figure 1) Therefore an augmentation deci-sion engine similar to those used in [25] [33] [49] and exem-plary decision making flow presented in this paper (discussedin part VI-C) to determine the mobile augmentation feasibilityis needed to amend the CMA performance and reliabilityDuring augmentation process the local and native applicationstate stack needs synchronization to ensure integrity betweennative and remote data Upon successful outsourcing remoteresults need to be returned and integrated to the source mobiledevice Thus the lsquoSynchronizerrsquo component needs to performin typical CMA approaches (see Figure 1)

Although CMA approaches can empower mobile process-ing and storage capabilities several disadvantages such asapplication development complexity and unauthorized accessto remote data demand a systematized plenary solutionPerformance of the CMA systems is highly influenced byvarious challenges and issues of wireless networking andcloud computing technologies CMA researchers require a

high performance elastic robust reliable and foreseeablecommunication throughput between mobile nodes and cloudservers which is not yet realized despite of remarkable effortsand achievements of communication and networking societiesCurrent shortcomings and deficiencies of wireless communi-cation and networking especially seamless connectivity andmobility high performance communication throughput pro-visioning and wireless data interception discourage systemanalysts engineers developers and entrepreneurs from de-ploying CMA-enabled mobile applications due to the high riskof system malfunction and user experience degradation

Moreover CMA systems require accurate estimation mech-anisms to predict the overall time and energy consumptionof communication and computation tasks while exploitingclouds Such estimation is a challenging task consideringhuge infrastructuresrsquo performance diversity [50] and policyheterogeneity [51] of cloud services in intermittent wirelessenvironment Despite of blooming efforts endeavoring to ana-lyze and comprehend the cloud computing model and behavior[52]ndash[55] CMA solutions are still unable to accurately fore-see required time and energy of exploiting cloud resourcesto execute intensive applications Additionally sundry cloudchallenges especially live VM migration infrastructureandplatform heterogeneity efficient allocation of clouds to tasksQoS management security privacy and trust in cloud increasesystem complexity and decrease successful CMA systemsadoption

Among limited studies of the domain [19] and [56] surveyremote execution and application offloading algorithms withfocus on how task offloading is performed in various effortsFernando et al [57] and Dinh et al [58] sought to explain theconvergence of mobile and cloud computing and distinguishit from the earlier domains such as cloud and grid computing[59] The authors describe issues particularly mobile applica-tion offloading privacy and security context awareness anddata management Sanaei Abolfazli Gani and Buyya [51]present a comprehensive survey on MCC with major focuson heterogeneity The authors describe the challenges andopportunities imposed by heterogeneity and identify hardwareplatform feature API and network as the roots of MCC

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 3

TABLE IL IST OF ACRONYMS AND CORRESPONDINGFULL FORMS

Acronym Full form

2D 2 Dimensional

2G 2nd Generation

3D 3 Dimensional

3G 3rd Generation

API Application Programming Interface

App Application (mobile application)

ARM Advanced RISC Machines

CMA Cloud-based Mobile Augmentation

CPU Central Processing Unit

DSL Domain Specific Language

DVMS Dynamic VM Synthesis

FTP File Transfer Protocol

GPU Graphics Processing Unit

GUI Graphical User Interface

IO InputOutput

IaaS Infrastructure as a Service

IP Internet Protocol

IP TV Internet Protocol Television

iSCSI Internet Small Computer System Interface

MCC Mobile Cloud Computing

MNO Mobile Network Operator

OS Operating System

P2P Peer-to-Peer

PC Personal Computer

QoS Quality of Service

RampD Research and Development

RAM Random Access Memory

RISC Reduced Instruction Set Computing

RPC Remote Procedure Call

SAL Service Abstraction Layer

SLA Service Level Agreement

TCP Transmission Control Protocol

UDDI Universal Description Discovery and Integration

UI User Interface

VM Virtual Machine

WAN Wide Area Network

Wi-Fi Wireless Fidelity

heterogeneity They explain major heterogeneity handlingap-proaches particularly virtualization service orientedarchi-tecture and semantic technology However the computingperformance distance elasticity availability reliability andmulti-tenancy of remote resources are marginally discussedin these studies that necessitate further research to explainthe impact of remote resources on augmentation process andhighlight paradigm shift from the unreliable surrogates to

reliable clouds

In this paper we survey the state-of-the-art mobile augmen-tation efforts that employ cloud computing infrastructures toenhance computing capabilities of resource-constraint mobiledevices especially smartphones To the best of our knowledgethis is the first effort that studies the impacts of cloud-basedcomputing resources on mobile augmentation process We dif-ferentiate augmentation from similar concepts of load sharingand remote execution and present augmentation motivationWe review efforts that endeavor to mitigate the mobile devicesrsquoshortcomings and classify them as hardware and software todevise a taxonomy The impacts of CMA in mobile comput-ing are presented The characteristics of cloud-based remoteresources and their role in CMA effectiveness are studied andclassified into four groups namely distant immobile cloudsproximate immobile computing entities proximate mobilecomputing entities and hybrid based on their mobility andphysical location traits Furthermore the state-of-the-art CMAmodels are reviewed and taxonomized into four classes ofdistant fixed proximate fixed proximate mobile and hybridaccording to our cloud-based resource classification Factorsimpact on the CMA adaptability are identified and described asaugmentation environment user preferences and requirementsmobile devices cloud servers and contents Five major metricsthat limit the performance of CMA approaches are studied Asample flowchart of decision making engines for imminentCMA solutions is presented and several open challenges arediscussed as the future research directions Such survey isbeneficial to the communication and networking communitiesbecause comprehending CMA process and current deploy-ment challenges are beneficial in modifying the fundamentalnetworking infrastructures to optimize the CMA process Inthis paper we use the terms mobile devices and smartphonesinterchangeably with similar notion Table I shows the listofacronyms used in the paper

The remainder of this paper is organized as follows SectionII introduces mobile computation augmentation presents itsmotivation and describes the taxonomy of mobile augmenta-tion types The impacts of CMA on mobile computing arepresented in Section III Section IV presents the analysisand taxonomy of varied cloud-based augmentation resourcesComprehensive survey of the state-of-the-art CMA approachesis presented and taxonomy is devised in Section V Wediscuss the CMA decision making and limitation factors andillustrate CMA feasibility in Section VI Finally open researchchallenges are presented in Section VII and paper is concludedin Section VIII

II M OBILE COMPUTATION AUGMENTATION

In this Section we present a definition on mobile computingaugmentation based on the available definitions on the relevantconcepts particularly remote execution [5] and cyber foraging[6] Additionally the motivation for performing mobile com-putation augmentation is described and taxonomy of mobileaugmentation types is presented

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 4

TABLE IIINITIAL FEATURES OFMOBILE EMPOWERMENTAPPROACHES

Approach Architecture Client Load Migration Partitioning Server Mobility

Load Sharing Client-Server Entire Task Entire task NA Server NA

RemoteClient-Server Entire Task Entirepartial Static

Server NoExecution desktop

Cyber Client-ServerEntire Task Entirepartial Dynamic

Surrogates NoForaging Peer-to-Peer

Mobile Varies eg

Nil

Entirepartial Static amp Server YesComputation Client-server Entirepartial Nil migration dynamic surrogateAugmentation P2P Adhoc (Use remote ampmobile

collaborative services)

A Definition

Indeed empowering computation capabilities of mobiledevices is not a new concept and there have been differentapproaches to achieve this goal including load sharing [4]remote execution [5] cyber foraging [6] and computationoffloading [60] [61] that are described as follows We haveanalyzed them and summarized the analysis results in TableII Results in this Table are extracted from the early efforts ineach category which are already deviated from their originalcharacteristics due to the research achievements

bull Load Sharing Othman and Hailesrsquo work [4] in 1998can be considered as one of the earliest efforts to conservenative resources of mobile devices using a software approachThe main idea is inspired from the concept of load balancingin distributed computing that is ldquoa strategy which attemptsto share loads in a distributed system without attemptingto equalize its loadrdquo [4] This approach migrates the wholecomputation job for remote execution It considers severalmetrics such as job size available bandwidth and result sizeto identify if the load balancing and transferring the job tothe remote computer can save energy However they need tosend the task and data to the nearest base station and wait forthe results to return The base station is responsible to findappropriate server to run the job and forward the results backto the mobile device Moreover computing server is a fixedcomputer and there is no provision for user and code mobilityat run time

bull Remote ExecutionThe concept of remote execution formobile clients emerged in 90s and several researchers [5][62]ndash[65] endeavor to enable mobile computers to performingremote computation and data storage to conserve their scarcenative resources and battery In 1998 [5] feasibility of remoteexecution concept on mobile computers particularly laptopswas investigated The authors report that remote executioncansave energy if the remote processing cost is lower than localexecution Remote execution involves migrating computingtasks from the mobile device to the server prior the executionThe server performs the task and sends back the results tothe mobile device However difference between computationpower of client and server is not a metric of decision makingin this method Moreover the whole task needs to be migratedto the remote server prior the execution which is an expensive

effort It also neglects the impact of environment characteris-tics on the remote execution outcome Static decision makingis another shortcoming of this proposal

bull Cyber Foraging Satyanarayana in 2001 [6] furtherdeveloped the remote execution idea by considering dynamismin remote execution process The author defined cyber forag-ing as the process ldquoto dynamically augment the computingresources of a wireless mobile computer by exploiting wiredhardware infrastructurerdquo Resources in cyber foraging arestationary computers or servers in public places mdashconnectedto wired Internet and power cablemdashthat are willing to performintensive computation on behalf of the resource-constraintmobile devices in vicinity

However load sharing remote execution and cyber forag-ing approaches assume that the whole computing task is storedin the device and hence it requires the intensive code and datato be identified and partitioned for offloading mdasheither stati-cally prior the execution or dynamically at runtime mdashwhichimpose large overhead on resource-poor mobile device [19]Moreover as Kumar et al [66] explain for each mobileuser that runs the intensive application the whole offloadingprocess must be repeated including decision making processin the device and transferring the heavy components and largedata to the network Due to slight differences among theseconcepts researchers use the terms lsquoremote executionrsquo lsquocyberforagingrsquo and lsquocomputation offloadingrsquo interchangeably in theliterature with similar principle and notion

Nevertheless researchers in recent activities [36] [42] [45][46] aim to enhance computing capabilities of mobile devicesin a slightly different manner They assume to store theintensive code and data outside the device and keep the rest inthe mobile device instead of storing the whole task mdashincludingboth lightweight and intensive code and data mdashin the mobiledevice Therefore the overhead of identifying partitioningand migrating the resource-intensive task is mitigated energyis saved and storage problem is alleviated in mobile devicesMoreover storing intensive components outside the devicein a publicly accessible storage can increase their reusabilityand enable more than one user to leverage their computationservices Therefore we coin the termmobile computationaugmentationas the wider phrase that subsumes load sharingremote execution cyber foraging and other approaches that

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 5

augment computing capabilities of mobile devicesbull Mobile Computation Augmentation Mobile computa-

tion augmentation or augmentation in brief is the processofincreasing enhancing and optimizing computing capabilitiesof mobile devices by leveraging varied feasible approacheshardware and software Mobile device is any non-stationarybattery-operating computing entity able to interact with end-user and execute transactions store data and communicatewith the environment using wireless technologies and variedsensors Smartphone Tablet handheldwearable computingdevices and vehicle mount computers are mobile device in-stances Approaches that can augment mobile devices includehardware and software Hardware approach involves manu-facturing high-end physical components particularly CPUmemory storage and battery Software approaches can bemdashbut are not limited to mdashcomputation offloading remote datastorage wireless communication resource-aware computingfidelity adaptation and remote service request (eg contextacquisition)

Augmentation approaches can increase computing capabil-ities of mobile devices and conserve energy They can beleveraged in three main categories of applications namely(i)computing-intensive software such as speech recognition andnatural language processing (ii) data-intensive programs suchas enterprise applications and (iii) communication-intensiveapplications such as online video streaming applicationsTheaugmented mobile device is able to perform complex tasks thatcould not otherwise perform Hence the mobile applicationdevelopers do not take into account resource shortcomingsof mobile devices in developing mobile application and userswill not consider their devices weaknesses in utilizing variedcomplex applications

B Motivation

Mobile devices have recently gained momentous ground inseveral communities like governmental agencies enterprisessocial service providers (eg insurance Police fire depart-ments) healthcare education and engineering organizations[67] [68] However despite of significant improvement inmobilesrsquo computing capabilities still computing requirementsof mobile users especially enterprise users is not achieved

Several intrinsic deficiencies of mobile devices encumberfeasibility of intense mobile computing and motivate aug-mentation Leveraging augmentation approaches vision ofperforming intense mobile operations and control such asremote surgery on-site engineering and visionary scenariossimilar to the lost child and disaster relief described in [69]will become reality In this part we analyze and taxonomizesmartphonesrsquo deficits that can be alleviated by augmentationFigure 2 depicts our devised taxonomy

1) Processing PowerProcessing deficiencies of mobileclients due to slow processing speed and limited RAM is oneof the major challenges in mobile computing [69] Mobile de-vices are expected to have high processing capabilities similarto computing capabilities of desktop machines for performingcomputing-intensive tasks to enrich user experience Realizingsuch vision requires powerful processor being able to performlarge number of transactions in a short course of time

Large internal memoryRAM to store state stack of allrunning applications and background services is also lackingBeside local memory limitations memory leakage also inten-sifies memory restrains of mobile devices Memory leakageis the state of memory cells being unnecessarily occupied byrunning applications and services or those cells that are notreleased after utilization Garbage-collector-based languageslike Java in Android2 contribute to memory leakage due tofailed or delayed removal of unused objects from the memory[70] Androidrsquos kernel level transactions can also leak memoryin the absence of memory management mechanisms [70][71] Moreover inward deficiency and inefficient design andimplementation of mobile applications can also waste scarcememory of mobile devices Thus in the absence of requiredmemory applications are frequently paused or terminatedby the operating system leading to longer execution timeexcessive resource dissipation and ultimately mobile userexperience degradation

2) Energy ResourcesEnergy is the only non-replenishableresource in mobile devices that demands external resourcesto be replenished [72] [73] Currently energy requirement ofa mobile device is supplied via lithium-ion battery that lastsonly few hours if device is computationally engaged Batterycapacity is increasing at about 5 to 10 a year [74] [75]as battery cells are excessively dense [72] Moreover mo-bile device manufacturers endeavor to attain device lightnesscompactness and handiness which prevent exploitation ofbulky long-lasting batteries User safety is another concern thatconfines manufacturers to produce low capacity batteries [76]While explosion of a battery with few hundreds milliamperescapacity can jeopardize human life [77] explosion of a high-capacity battery can carry catastrophic consequences

Energy harvesting efforts [78]ndash[80] seek to replenish energyfrom renewable resources particularly human movement solarenergy and wireless radiation but these resources are mostlyintermittent and not available on-demand [81] For instancea sitting mobile user at night cannot have any external powersource in the absence of wall power and wireless radiationsMoreover researchers aim at reducing the energy overheadin different aspects of computing including hardware OSapplication and networking interface [82] [83] Effortsare di-rected to develop alternative energy resources such as nuclearbatteries that will likely last months or years [84] Howeversignificant deal of RampD is needed to fulfill ever-increasingenergy requirements of mobile users

Hence in the absence of long-spanning energy on mobiledevices alternative augmentation approaches play a vitalrolein maturing mobile and ubiquitous computing

3) Local Storage Drastic increasing in the number ofapplications and amount of digital contents such as picturessongs movies and home films [85] from one hand and limitedstorage of mobile devices from the other hand decelerateusability of mobile devices While PCs are able to locallystore huge amount of data smartphones are limited to fewgigabytes of space which are mostly occupied by systemfiles user applications and personal data Therefore frequent

2urlhttpwwwandroidcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 6

Fig 2 Taxonomy of Augmentation Motivation Intrinsic andnon-intrinsic mobile challenges motivate augmentation

storing updating and deleting data as well as uninstalling andreinstalling applications due to space limitation cause irksomeimpediments to mobile users [86] Additionally deliveringoffline usability which is one of the most important character-istics of contemporary applications remains an open challengesince mobile devices lack large local storage

4) Visualization CapabilitiesEffective data visualizationon small mobile devicesrsquo screen is a non-trivial task whencurrent screen manufacturing technologies and energy limita-tions do not allow significant size extensions without losingdevice handiness Currently smartphones like HTC One X3

and Samsung Galaxy Note II4 have the biggest screens at47 and 55 inches respectively however they are very smallcompared to PCs and notebooks

Therefore efficient data visualization in small smartphonesrsquoscreen necessitates software-based techniques similar totab-ular pages 3D objects multiple desktops switching betweenlandscape and portrait views (needs accelerometer) and verbalcommunication to virtually increase presentation area Re-cently computing-intensive mobile 3D display technologyispromising to noticeably mitigate the visualization deficitofcontemporary smartphones Glass-free auto-stereoscopicdis-plays [87] can present 3D data by exploiting binocular parallaxto offer a different view for each eye Taking advantagesof current and imminent software-based techniques besidenative tools especially tilting sensors significantly improve themobile visualization capabilities in the near future Howeverthese approaches are computation-intensive processes thatquickly drain battery [87] [88] A feasible alternative solutionto realize software-based content presentation approaches is toaugment smartphonesrsquo computing capabilities

5) Security Privacy and Data SafetyMobile end-usersare concerned about security and privacy of their personaldata banking records and online behaviors [89] The dramaticincrease in cybercrime and security threats within mobiledevices [90] cloud resources [91] and wireless transactionsmakes security and privacy more challenging than ever [92]Moreover performing complex cryptographic algorithms islikely infeasible because of computing deficiencies of mobile

3httpwwwhtccomwwwsmartphoneshtc-one-x4httpwwwsamsungcommyconsumermobile-devicesgalaxy-

notegalaxy-noteGT-N7100RWDXME

devices Securing files using pair of credentials is also lessrealistic in the absence of large keyboard

Data safety is another challenge of mobile devices becauseinformation stored inside the local storage of mobile devicesare susceptible to safety breaches due to high probability ofhardware malfunction physical damage stealing and loss

Amalgam of these problems and deficiencies in mobilecomputing stimulates researchers from academia and industryto exploit novel technologies and approaches to augmentcomputing capabilities of mobile devices which is subject ofthis study

C Mobile Augmentation Types Taxonomy

In this Section we analyze and classify augmentation ap-proaches into two major types of hardware and software Ourdevised taxonomy is depicted in Figure 3 and described asfollowsHardware The hardware approach aims to empower smart-phones by exploiting powerful resources particularly multi-core CPUs with high clock speed [93] large screen and long-lasting battery [84] [94] ARM5 and Samsung6 are major mo-bile processor manufacturers producing multi-core processorssuch as ARM Cortex-597 and Samsung Exynos 5 Octa core8

that perform in higher speed than a single core processors[93] However doubling the CPU clock speed approximatelyoctuples the device energy consumption [66]

Nevertheless augmentation via sophisticated hardware ishindered by several obstacles Firstly generating powerfulprocessor large storage and big screen decrease smartphonehandiness due to additional heat size and weight Secondlyconsidering the fact that utilizing long-lasting battery in smallmobile devices is not feasible with current technologies re-source enlargement contributes toward faster battery drainageand shorter battery life time Thirdly equipping mobile de-vices with high-end hardware noticeably increases their price

5httparmcom6httpsamsungcom7httpwwwarmcomproductsprocessorscortex-a50cortex-a57-

processorphp8httpwwwsamsungcomglobalbusinesssemiconductorminisiteExynos

blog CES 2013 SamsungMobilizes Possibility with Exynos5Octahtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 3: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 3

TABLE IL IST OF ACRONYMS AND CORRESPONDINGFULL FORMS

Acronym Full form

2D 2 Dimensional

2G 2nd Generation

3D 3 Dimensional

3G 3rd Generation

API Application Programming Interface

App Application (mobile application)

ARM Advanced RISC Machines

CMA Cloud-based Mobile Augmentation

CPU Central Processing Unit

DSL Domain Specific Language

DVMS Dynamic VM Synthesis

FTP File Transfer Protocol

GPU Graphics Processing Unit

GUI Graphical User Interface

IO InputOutput

IaaS Infrastructure as a Service

IP Internet Protocol

IP TV Internet Protocol Television

iSCSI Internet Small Computer System Interface

MCC Mobile Cloud Computing

MNO Mobile Network Operator

OS Operating System

P2P Peer-to-Peer

PC Personal Computer

QoS Quality of Service

RampD Research and Development

RAM Random Access Memory

RISC Reduced Instruction Set Computing

RPC Remote Procedure Call

SAL Service Abstraction Layer

SLA Service Level Agreement

TCP Transmission Control Protocol

UDDI Universal Description Discovery and Integration

UI User Interface

VM Virtual Machine

WAN Wide Area Network

Wi-Fi Wireless Fidelity

heterogeneity They explain major heterogeneity handlingap-proaches particularly virtualization service orientedarchi-tecture and semantic technology However the computingperformance distance elasticity availability reliability andmulti-tenancy of remote resources are marginally discussedin these studies that necessitate further research to explainthe impact of remote resources on augmentation process andhighlight paradigm shift from the unreliable surrogates to

reliable clouds

In this paper we survey the state-of-the-art mobile augmen-tation efforts that employ cloud computing infrastructures toenhance computing capabilities of resource-constraint mobiledevices especially smartphones To the best of our knowledgethis is the first effort that studies the impacts of cloud-basedcomputing resources on mobile augmentation process We dif-ferentiate augmentation from similar concepts of load sharingand remote execution and present augmentation motivationWe review efforts that endeavor to mitigate the mobile devicesrsquoshortcomings and classify them as hardware and software todevise a taxonomy The impacts of CMA in mobile comput-ing are presented The characteristics of cloud-based remoteresources and their role in CMA effectiveness are studied andclassified into four groups namely distant immobile cloudsproximate immobile computing entities proximate mobilecomputing entities and hybrid based on their mobility andphysical location traits Furthermore the state-of-the-art CMAmodels are reviewed and taxonomized into four classes ofdistant fixed proximate fixed proximate mobile and hybridaccording to our cloud-based resource classification Factorsimpact on the CMA adaptability are identified and described asaugmentation environment user preferences and requirementsmobile devices cloud servers and contents Five major metricsthat limit the performance of CMA approaches are studied Asample flowchart of decision making engines for imminentCMA solutions is presented and several open challenges arediscussed as the future research directions Such survey isbeneficial to the communication and networking communitiesbecause comprehending CMA process and current deploy-ment challenges are beneficial in modifying the fundamentalnetworking infrastructures to optimize the CMA process Inthis paper we use the terms mobile devices and smartphonesinterchangeably with similar notion Table I shows the listofacronyms used in the paper

The remainder of this paper is organized as follows SectionII introduces mobile computation augmentation presents itsmotivation and describes the taxonomy of mobile augmenta-tion types The impacts of CMA on mobile computing arepresented in Section III Section IV presents the analysisand taxonomy of varied cloud-based augmentation resourcesComprehensive survey of the state-of-the-art CMA approachesis presented and taxonomy is devised in Section V Wediscuss the CMA decision making and limitation factors andillustrate CMA feasibility in Section VI Finally open researchchallenges are presented in Section VII and paper is concludedin Section VIII

II M OBILE COMPUTATION AUGMENTATION

In this Section we present a definition on mobile computingaugmentation based on the available definitions on the relevantconcepts particularly remote execution [5] and cyber foraging[6] Additionally the motivation for performing mobile com-putation augmentation is described and taxonomy of mobileaugmentation types is presented

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 4

TABLE IIINITIAL FEATURES OFMOBILE EMPOWERMENTAPPROACHES

Approach Architecture Client Load Migration Partitioning Server Mobility

Load Sharing Client-Server Entire Task Entire task NA Server NA

RemoteClient-Server Entire Task Entirepartial Static

Server NoExecution desktop

Cyber Client-ServerEntire Task Entirepartial Dynamic

Surrogates NoForaging Peer-to-Peer

Mobile Varies eg

Nil

Entirepartial Static amp Server YesComputation Client-server Entirepartial Nil migration dynamic surrogateAugmentation P2P Adhoc (Use remote ampmobile

collaborative services)

A Definition

Indeed empowering computation capabilities of mobiledevices is not a new concept and there have been differentapproaches to achieve this goal including load sharing [4]remote execution [5] cyber foraging [6] and computationoffloading [60] [61] that are described as follows We haveanalyzed them and summarized the analysis results in TableII Results in this Table are extracted from the early efforts ineach category which are already deviated from their originalcharacteristics due to the research achievements

bull Load Sharing Othman and Hailesrsquo work [4] in 1998can be considered as one of the earliest efforts to conservenative resources of mobile devices using a software approachThe main idea is inspired from the concept of load balancingin distributed computing that is ldquoa strategy which attemptsto share loads in a distributed system without attemptingto equalize its loadrdquo [4] This approach migrates the wholecomputation job for remote execution It considers severalmetrics such as job size available bandwidth and result sizeto identify if the load balancing and transferring the job tothe remote computer can save energy However they need tosend the task and data to the nearest base station and wait forthe results to return The base station is responsible to findappropriate server to run the job and forward the results backto the mobile device Moreover computing server is a fixedcomputer and there is no provision for user and code mobilityat run time

bull Remote ExecutionThe concept of remote execution formobile clients emerged in 90s and several researchers [5][62]ndash[65] endeavor to enable mobile computers to performingremote computation and data storage to conserve their scarcenative resources and battery In 1998 [5] feasibility of remoteexecution concept on mobile computers particularly laptopswas investigated The authors report that remote executioncansave energy if the remote processing cost is lower than localexecution Remote execution involves migrating computingtasks from the mobile device to the server prior the executionThe server performs the task and sends back the results tothe mobile device However difference between computationpower of client and server is not a metric of decision makingin this method Moreover the whole task needs to be migratedto the remote server prior the execution which is an expensive

effort It also neglects the impact of environment characteris-tics on the remote execution outcome Static decision makingis another shortcoming of this proposal

bull Cyber Foraging Satyanarayana in 2001 [6] furtherdeveloped the remote execution idea by considering dynamismin remote execution process The author defined cyber forag-ing as the process ldquoto dynamically augment the computingresources of a wireless mobile computer by exploiting wiredhardware infrastructurerdquo Resources in cyber foraging arestationary computers or servers in public places mdashconnectedto wired Internet and power cablemdashthat are willing to performintensive computation on behalf of the resource-constraintmobile devices in vicinity

However load sharing remote execution and cyber forag-ing approaches assume that the whole computing task is storedin the device and hence it requires the intensive code and datato be identified and partitioned for offloading mdasheither stati-cally prior the execution or dynamically at runtime mdashwhichimpose large overhead on resource-poor mobile device [19]Moreover as Kumar et al [66] explain for each mobileuser that runs the intensive application the whole offloadingprocess must be repeated including decision making processin the device and transferring the heavy components and largedata to the network Due to slight differences among theseconcepts researchers use the terms lsquoremote executionrsquo lsquocyberforagingrsquo and lsquocomputation offloadingrsquo interchangeably in theliterature with similar principle and notion

Nevertheless researchers in recent activities [36] [42] [45][46] aim to enhance computing capabilities of mobile devicesin a slightly different manner They assume to store theintensive code and data outside the device and keep the rest inthe mobile device instead of storing the whole task mdashincludingboth lightweight and intensive code and data mdashin the mobiledevice Therefore the overhead of identifying partitioningand migrating the resource-intensive task is mitigated energyis saved and storage problem is alleviated in mobile devicesMoreover storing intensive components outside the devicein a publicly accessible storage can increase their reusabilityand enable more than one user to leverage their computationservices Therefore we coin the termmobile computationaugmentationas the wider phrase that subsumes load sharingremote execution cyber foraging and other approaches that

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 5

augment computing capabilities of mobile devicesbull Mobile Computation Augmentation Mobile computa-

tion augmentation or augmentation in brief is the processofincreasing enhancing and optimizing computing capabilitiesof mobile devices by leveraging varied feasible approacheshardware and software Mobile device is any non-stationarybattery-operating computing entity able to interact with end-user and execute transactions store data and communicatewith the environment using wireless technologies and variedsensors Smartphone Tablet handheldwearable computingdevices and vehicle mount computers are mobile device in-stances Approaches that can augment mobile devices includehardware and software Hardware approach involves manu-facturing high-end physical components particularly CPUmemory storage and battery Software approaches can bemdashbut are not limited to mdashcomputation offloading remote datastorage wireless communication resource-aware computingfidelity adaptation and remote service request (eg contextacquisition)

Augmentation approaches can increase computing capabil-ities of mobile devices and conserve energy They can beleveraged in three main categories of applications namely(i)computing-intensive software such as speech recognition andnatural language processing (ii) data-intensive programs suchas enterprise applications and (iii) communication-intensiveapplications such as online video streaming applicationsTheaugmented mobile device is able to perform complex tasks thatcould not otherwise perform Hence the mobile applicationdevelopers do not take into account resource shortcomingsof mobile devices in developing mobile application and userswill not consider their devices weaknesses in utilizing variedcomplex applications

B Motivation

Mobile devices have recently gained momentous ground inseveral communities like governmental agencies enterprisessocial service providers (eg insurance Police fire depart-ments) healthcare education and engineering organizations[67] [68] However despite of significant improvement inmobilesrsquo computing capabilities still computing requirementsof mobile users especially enterprise users is not achieved

Several intrinsic deficiencies of mobile devices encumberfeasibility of intense mobile computing and motivate aug-mentation Leveraging augmentation approaches vision ofperforming intense mobile operations and control such asremote surgery on-site engineering and visionary scenariossimilar to the lost child and disaster relief described in [69]will become reality In this part we analyze and taxonomizesmartphonesrsquo deficits that can be alleviated by augmentationFigure 2 depicts our devised taxonomy

1) Processing PowerProcessing deficiencies of mobileclients due to slow processing speed and limited RAM is oneof the major challenges in mobile computing [69] Mobile de-vices are expected to have high processing capabilities similarto computing capabilities of desktop machines for performingcomputing-intensive tasks to enrich user experience Realizingsuch vision requires powerful processor being able to performlarge number of transactions in a short course of time

Large internal memoryRAM to store state stack of allrunning applications and background services is also lackingBeside local memory limitations memory leakage also inten-sifies memory restrains of mobile devices Memory leakageis the state of memory cells being unnecessarily occupied byrunning applications and services or those cells that are notreleased after utilization Garbage-collector-based languageslike Java in Android2 contribute to memory leakage due tofailed or delayed removal of unused objects from the memory[70] Androidrsquos kernel level transactions can also leak memoryin the absence of memory management mechanisms [70][71] Moreover inward deficiency and inefficient design andimplementation of mobile applications can also waste scarcememory of mobile devices Thus in the absence of requiredmemory applications are frequently paused or terminatedby the operating system leading to longer execution timeexcessive resource dissipation and ultimately mobile userexperience degradation

2) Energy ResourcesEnergy is the only non-replenishableresource in mobile devices that demands external resourcesto be replenished [72] [73] Currently energy requirement ofa mobile device is supplied via lithium-ion battery that lastsonly few hours if device is computationally engaged Batterycapacity is increasing at about 5 to 10 a year [74] [75]as battery cells are excessively dense [72] Moreover mo-bile device manufacturers endeavor to attain device lightnesscompactness and handiness which prevent exploitation ofbulky long-lasting batteries User safety is another concern thatconfines manufacturers to produce low capacity batteries [76]While explosion of a battery with few hundreds milliamperescapacity can jeopardize human life [77] explosion of a high-capacity battery can carry catastrophic consequences

Energy harvesting efforts [78]ndash[80] seek to replenish energyfrom renewable resources particularly human movement solarenergy and wireless radiation but these resources are mostlyintermittent and not available on-demand [81] For instancea sitting mobile user at night cannot have any external powersource in the absence of wall power and wireless radiationsMoreover researchers aim at reducing the energy overheadin different aspects of computing including hardware OSapplication and networking interface [82] [83] Effortsare di-rected to develop alternative energy resources such as nuclearbatteries that will likely last months or years [84] Howeversignificant deal of RampD is needed to fulfill ever-increasingenergy requirements of mobile users

Hence in the absence of long-spanning energy on mobiledevices alternative augmentation approaches play a vitalrolein maturing mobile and ubiquitous computing

3) Local Storage Drastic increasing in the number ofapplications and amount of digital contents such as picturessongs movies and home films [85] from one hand and limitedstorage of mobile devices from the other hand decelerateusability of mobile devices While PCs are able to locallystore huge amount of data smartphones are limited to fewgigabytes of space which are mostly occupied by systemfiles user applications and personal data Therefore frequent

2urlhttpwwwandroidcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 6

Fig 2 Taxonomy of Augmentation Motivation Intrinsic andnon-intrinsic mobile challenges motivate augmentation

storing updating and deleting data as well as uninstalling andreinstalling applications due to space limitation cause irksomeimpediments to mobile users [86] Additionally deliveringoffline usability which is one of the most important character-istics of contemporary applications remains an open challengesince mobile devices lack large local storage

4) Visualization CapabilitiesEffective data visualizationon small mobile devicesrsquo screen is a non-trivial task whencurrent screen manufacturing technologies and energy limita-tions do not allow significant size extensions without losingdevice handiness Currently smartphones like HTC One X3

and Samsung Galaxy Note II4 have the biggest screens at47 and 55 inches respectively however they are very smallcompared to PCs and notebooks

Therefore efficient data visualization in small smartphonesrsquoscreen necessitates software-based techniques similar totab-ular pages 3D objects multiple desktops switching betweenlandscape and portrait views (needs accelerometer) and verbalcommunication to virtually increase presentation area Re-cently computing-intensive mobile 3D display technologyispromising to noticeably mitigate the visualization deficitofcontemporary smartphones Glass-free auto-stereoscopicdis-plays [87] can present 3D data by exploiting binocular parallaxto offer a different view for each eye Taking advantagesof current and imminent software-based techniques besidenative tools especially tilting sensors significantly improve themobile visualization capabilities in the near future Howeverthese approaches are computation-intensive processes thatquickly drain battery [87] [88] A feasible alternative solutionto realize software-based content presentation approaches is toaugment smartphonesrsquo computing capabilities

5) Security Privacy and Data SafetyMobile end-usersare concerned about security and privacy of their personaldata banking records and online behaviors [89] The dramaticincrease in cybercrime and security threats within mobiledevices [90] cloud resources [91] and wireless transactionsmakes security and privacy more challenging than ever [92]Moreover performing complex cryptographic algorithms islikely infeasible because of computing deficiencies of mobile

3httpwwwhtccomwwwsmartphoneshtc-one-x4httpwwwsamsungcommyconsumermobile-devicesgalaxy-

notegalaxy-noteGT-N7100RWDXME

devices Securing files using pair of credentials is also lessrealistic in the absence of large keyboard

Data safety is another challenge of mobile devices becauseinformation stored inside the local storage of mobile devicesare susceptible to safety breaches due to high probability ofhardware malfunction physical damage stealing and loss

Amalgam of these problems and deficiencies in mobilecomputing stimulates researchers from academia and industryto exploit novel technologies and approaches to augmentcomputing capabilities of mobile devices which is subject ofthis study

C Mobile Augmentation Types Taxonomy

In this Section we analyze and classify augmentation ap-proaches into two major types of hardware and software Ourdevised taxonomy is depicted in Figure 3 and described asfollowsHardware The hardware approach aims to empower smart-phones by exploiting powerful resources particularly multi-core CPUs with high clock speed [93] large screen and long-lasting battery [84] [94] ARM5 and Samsung6 are major mo-bile processor manufacturers producing multi-core processorssuch as ARM Cortex-597 and Samsung Exynos 5 Octa core8

that perform in higher speed than a single core processors[93] However doubling the CPU clock speed approximatelyoctuples the device energy consumption [66]

Nevertheless augmentation via sophisticated hardware ishindered by several obstacles Firstly generating powerfulprocessor large storage and big screen decrease smartphonehandiness due to additional heat size and weight Secondlyconsidering the fact that utilizing long-lasting battery in smallmobile devices is not feasible with current technologies re-source enlargement contributes toward faster battery drainageand shorter battery life time Thirdly equipping mobile de-vices with high-end hardware noticeably increases their price

5httparmcom6httpsamsungcom7httpwwwarmcomproductsprocessorscortex-a50cortex-a57-

processorphp8httpwwwsamsungcomglobalbusinesssemiconductorminisiteExynos

blog CES 2013 SamsungMobilizes Possibility with Exynos5Octahtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 4: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 4

TABLE IIINITIAL FEATURES OFMOBILE EMPOWERMENTAPPROACHES

Approach Architecture Client Load Migration Partitioning Server Mobility

Load Sharing Client-Server Entire Task Entire task NA Server NA

RemoteClient-Server Entire Task Entirepartial Static

Server NoExecution desktop

Cyber Client-ServerEntire Task Entirepartial Dynamic

Surrogates NoForaging Peer-to-Peer

Mobile Varies eg

Nil

Entirepartial Static amp Server YesComputation Client-server Entirepartial Nil migration dynamic surrogateAugmentation P2P Adhoc (Use remote ampmobile

collaborative services)

A Definition

Indeed empowering computation capabilities of mobiledevices is not a new concept and there have been differentapproaches to achieve this goal including load sharing [4]remote execution [5] cyber foraging [6] and computationoffloading [60] [61] that are described as follows We haveanalyzed them and summarized the analysis results in TableII Results in this Table are extracted from the early efforts ineach category which are already deviated from their originalcharacteristics due to the research achievements

bull Load Sharing Othman and Hailesrsquo work [4] in 1998can be considered as one of the earliest efforts to conservenative resources of mobile devices using a software approachThe main idea is inspired from the concept of load balancingin distributed computing that is ldquoa strategy which attemptsto share loads in a distributed system without attemptingto equalize its loadrdquo [4] This approach migrates the wholecomputation job for remote execution It considers severalmetrics such as job size available bandwidth and result sizeto identify if the load balancing and transferring the job tothe remote computer can save energy However they need tosend the task and data to the nearest base station and wait forthe results to return The base station is responsible to findappropriate server to run the job and forward the results backto the mobile device Moreover computing server is a fixedcomputer and there is no provision for user and code mobilityat run time

bull Remote ExecutionThe concept of remote execution formobile clients emerged in 90s and several researchers [5][62]ndash[65] endeavor to enable mobile computers to performingremote computation and data storage to conserve their scarcenative resources and battery In 1998 [5] feasibility of remoteexecution concept on mobile computers particularly laptopswas investigated The authors report that remote executioncansave energy if the remote processing cost is lower than localexecution Remote execution involves migrating computingtasks from the mobile device to the server prior the executionThe server performs the task and sends back the results tothe mobile device However difference between computationpower of client and server is not a metric of decision makingin this method Moreover the whole task needs to be migratedto the remote server prior the execution which is an expensive

effort It also neglects the impact of environment characteris-tics on the remote execution outcome Static decision makingis another shortcoming of this proposal

bull Cyber Foraging Satyanarayana in 2001 [6] furtherdeveloped the remote execution idea by considering dynamismin remote execution process The author defined cyber forag-ing as the process ldquoto dynamically augment the computingresources of a wireless mobile computer by exploiting wiredhardware infrastructurerdquo Resources in cyber foraging arestationary computers or servers in public places mdashconnectedto wired Internet and power cablemdashthat are willing to performintensive computation on behalf of the resource-constraintmobile devices in vicinity

However load sharing remote execution and cyber forag-ing approaches assume that the whole computing task is storedin the device and hence it requires the intensive code and datato be identified and partitioned for offloading mdasheither stati-cally prior the execution or dynamically at runtime mdashwhichimpose large overhead on resource-poor mobile device [19]Moreover as Kumar et al [66] explain for each mobileuser that runs the intensive application the whole offloadingprocess must be repeated including decision making processin the device and transferring the heavy components and largedata to the network Due to slight differences among theseconcepts researchers use the terms lsquoremote executionrsquo lsquocyberforagingrsquo and lsquocomputation offloadingrsquo interchangeably in theliterature with similar principle and notion

Nevertheless researchers in recent activities [36] [42] [45][46] aim to enhance computing capabilities of mobile devicesin a slightly different manner They assume to store theintensive code and data outside the device and keep the rest inthe mobile device instead of storing the whole task mdashincludingboth lightweight and intensive code and data mdashin the mobiledevice Therefore the overhead of identifying partitioningand migrating the resource-intensive task is mitigated energyis saved and storage problem is alleviated in mobile devicesMoreover storing intensive components outside the devicein a publicly accessible storage can increase their reusabilityand enable more than one user to leverage their computationservices Therefore we coin the termmobile computationaugmentationas the wider phrase that subsumes load sharingremote execution cyber foraging and other approaches that

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 5

augment computing capabilities of mobile devicesbull Mobile Computation Augmentation Mobile computa-

tion augmentation or augmentation in brief is the processofincreasing enhancing and optimizing computing capabilitiesof mobile devices by leveraging varied feasible approacheshardware and software Mobile device is any non-stationarybattery-operating computing entity able to interact with end-user and execute transactions store data and communicatewith the environment using wireless technologies and variedsensors Smartphone Tablet handheldwearable computingdevices and vehicle mount computers are mobile device in-stances Approaches that can augment mobile devices includehardware and software Hardware approach involves manu-facturing high-end physical components particularly CPUmemory storage and battery Software approaches can bemdashbut are not limited to mdashcomputation offloading remote datastorage wireless communication resource-aware computingfidelity adaptation and remote service request (eg contextacquisition)

Augmentation approaches can increase computing capabil-ities of mobile devices and conserve energy They can beleveraged in three main categories of applications namely(i)computing-intensive software such as speech recognition andnatural language processing (ii) data-intensive programs suchas enterprise applications and (iii) communication-intensiveapplications such as online video streaming applicationsTheaugmented mobile device is able to perform complex tasks thatcould not otherwise perform Hence the mobile applicationdevelopers do not take into account resource shortcomingsof mobile devices in developing mobile application and userswill not consider their devices weaknesses in utilizing variedcomplex applications

B Motivation

Mobile devices have recently gained momentous ground inseveral communities like governmental agencies enterprisessocial service providers (eg insurance Police fire depart-ments) healthcare education and engineering organizations[67] [68] However despite of significant improvement inmobilesrsquo computing capabilities still computing requirementsof mobile users especially enterprise users is not achieved

Several intrinsic deficiencies of mobile devices encumberfeasibility of intense mobile computing and motivate aug-mentation Leveraging augmentation approaches vision ofperforming intense mobile operations and control such asremote surgery on-site engineering and visionary scenariossimilar to the lost child and disaster relief described in [69]will become reality In this part we analyze and taxonomizesmartphonesrsquo deficits that can be alleviated by augmentationFigure 2 depicts our devised taxonomy

1) Processing PowerProcessing deficiencies of mobileclients due to slow processing speed and limited RAM is oneof the major challenges in mobile computing [69] Mobile de-vices are expected to have high processing capabilities similarto computing capabilities of desktop machines for performingcomputing-intensive tasks to enrich user experience Realizingsuch vision requires powerful processor being able to performlarge number of transactions in a short course of time

Large internal memoryRAM to store state stack of allrunning applications and background services is also lackingBeside local memory limitations memory leakage also inten-sifies memory restrains of mobile devices Memory leakageis the state of memory cells being unnecessarily occupied byrunning applications and services or those cells that are notreleased after utilization Garbage-collector-based languageslike Java in Android2 contribute to memory leakage due tofailed or delayed removal of unused objects from the memory[70] Androidrsquos kernel level transactions can also leak memoryin the absence of memory management mechanisms [70][71] Moreover inward deficiency and inefficient design andimplementation of mobile applications can also waste scarcememory of mobile devices Thus in the absence of requiredmemory applications are frequently paused or terminatedby the operating system leading to longer execution timeexcessive resource dissipation and ultimately mobile userexperience degradation

2) Energy ResourcesEnergy is the only non-replenishableresource in mobile devices that demands external resourcesto be replenished [72] [73] Currently energy requirement ofa mobile device is supplied via lithium-ion battery that lastsonly few hours if device is computationally engaged Batterycapacity is increasing at about 5 to 10 a year [74] [75]as battery cells are excessively dense [72] Moreover mo-bile device manufacturers endeavor to attain device lightnesscompactness and handiness which prevent exploitation ofbulky long-lasting batteries User safety is another concern thatconfines manufacturers to produce low capacity batteries [76]While explosion of a battery with few hundreds milliamperescapacity can jeopardize human life [77] explosion of a high-capacity battery can carry catastrophic consequences

Energy harvesting efforts [78]ndash[80] seek to replenish energyfrom renewable resources particularly human movement solarenergy and wireless radiation but these resources are mostlyintermittent and not available on-demand [81] For instancea sitting mobile user at night cannot have any external powersource in the absence of wall power and wireless radiationsMoreover researchers aim at reducing the energy overheadin different aspects of computing including hardware OSapplication and networking interface [82] [83] Effortsare di-rected to develop alternative energy resources such as nuclearbatteries that will likely last months or years [84] Howeversignificant deal of RampD is needed to fulfill ever-increasingenergy requirements of mobile users

Hence in the absence of long-spanning energy on mobiledevices alternative augmentation approaches play a vitalrolein maturing mobile and ubiquitous computing

3) Local Storage Drastic increasing in the number ofapplications and amount of digital contents such as picturessongs movies and home films [85] from one hand and limitedstorage of mobile devices from the other hand decelerateusability of mobile devices While PCs are able to locallystore huge amount of data smartphones are limited to fewgigabytes of space which are mostly occupied by systemfiles user applications and personal data Therefore frequent

2urlhttpwwwandroidcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 6

Fig 2 Taxonomy of Augmentation Motivation Intrinsic andnon-intrinsic mobile challenges motivate augmentation

storing updating and deleting data as well as uninstalling andreinstalling applications due to space limitation cause irksomeimpediments to mobile users [86] Additionally deliveringoffline usability which is one of the most important character-istics of contemporary applications remains an open challengesince mobile devices lack large local storage

4) Visualization CapabilitiesEffective data visualizationon small mobile devicesrsquo screen is a non-trivial task whencurrent screen manufacturing technologies and energy limita-tions do not allow significant size extensions without losingdevice handiness Currently smartphones like HTC One X3

and Samsung Galaxy Note II4 have the biggest screens at47 and 55 inches respectively however they are very smallcompared to PCs and notebooks

Therefore efficient data visualization in small smartphonesrsquoscreen necessitates software-based techniques similar totab-ular pages 3D objects multiple desktops switching betweenlandscape and portrait views (needs accelerometer) and verbalcommunication to virtually increase presentation area Re-cently computing-intensive mobile 3D display technologyispromising to noticeably mitigate the visualization deficitofcontemporary smartphones Glass-free auto-stereoscopicdis-plays [87] can present 3D data by exploiting binocular parallaxto offer a different view for each eye Taking advantagesof current and imminent software-based techniques besidenative tools especially tilting sensors significantly improve themobile visualization capabilities in the near future Howeverthese approaches are computation-intensive processes thatquickly drain battery [87] [88] A feasible alternative solutionto realize software-based content presentation approaches is toaugment smartphonesrsquo computing capabilities

5) Security Privacy and Data SafetyMobile end-usersare concerned about security and privacy of their personaldata banking records and online behaviors [89] The dramaticincrease in cybercrime and security threats within mobiledevices [90] cloud resources [91] and wireless transactionsmakes security and privacy more challenging than ever [92]Moreover performing complex cryptographic algorithms islikely infeasible because of computing deficiencies of mobile

3httpwwwhtccomwwwsmartphoneshtc-one-x4httpwwwsamsungcommyconsumermobile-devicesgalaxy-

notegalaxy-noteGT-N7100RWDXME

devices Securing files using pair of credentials is also lessrealistic in the absence of large keyboard

Data safety is another challenge of mobile devices becauseinformation stored inside the local storage of mobile devicesare susceptible to safety breaches due to high probability ofhardware malfunction physical damage stealing and loss

Amalgam of these problems and deficiencies in mobilecomputing stimulates researchers from academia and industryto exploit novel technologies and approaches to augmentcomputing capabilities of mobile devices which is subject ofthis study

C Mobile Augmentation Types Taxonomy

In this Section we analyze and classify augmentation ap-proaches into two major types of hardware and software Ourdevised taxonomy is depicted in Figure 3 and described asfollowsHardware The hardware approach aims to empower smart-phones by exploiting powerful resources particularly multi-core CPUs with high clock speed [93] large screen and long-lasting battery [84] [94] ARM5 and Samsung6 are major mo-bile processor manufacturers producing multi-core processorssuch as ARM Cortex-597 and Samsung Exynos 5 Octa core8

that perform in higher speed than a single core processors[93] However doubling the CPU clock speed approximatelyoctuples the device energy consumption [66]

Nevertheless augmentation via sophisticated hardware ishindered by several obstacles Firstly generating powerfulprocessor large storage and big screen decrease smartphonehandiness due to additional heat size and weight Secondlyconsidering the fact that utilizing long-lasting battery in smallmobile devices is not feasible with current technologies re-source enlargement contributes toward faster battery drainageand shorter battery life time Thirdly equipping mobile de-vices with high-end hardware noticeably increases their price

5httparmcom6httpsamsungcom7httpwwwarmcomproductsprocessorscortex-a50cortex-a57-

processorphp8httpwwwsamsungcomglobalbusinesssemiconductorminisiteExynos

blog CES 2013 SamsungMobilizes Possibility with Exynos5Octahtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 5: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 5

augment computing capabilities of mobile devicesbull Mobile Computation Augmentation Mobile computa-

tion augmentation or augmentation in brief is the processofincreasing enhancing and optimizing computing capabilitiesof mobile devices by leveraging varied feasible approacheshardware and software Mobile device is any non-stationarybattery-operating computing entity able to interact with end-user and execute transactions store data and communicatewith the environment using wireless technologies and variedsensors Smartphone Tablet handheldwearable computingdevices and vehicle mount computers are mobile device in-stances Approaches that can augment mobile devices includehardware and software Hardware approach involves manu-facturing high-end physical components particularly CPUmemory storage and battery Software approaches can bemdashbut are not limited to mdashcomputation offloading remote datastorage wireless communication resource-aware computingfidelity adaptation and remote service request (eg contextacquisition)

Augmentation approaches can increase computing capabil-ities of mobile devices and conserve energy They can beleveraged in three main categories of applications namely(i)computing-intensive software such as speech recognition andnatural language processing (ii) data-intensive programs suchas enterprise applications and (iii) communication-intensiveapplications such as online video streaming applicationsTheaugmented mobile device is able to perform complex tasks thatcould not otherwise perform Hence the mobile applicationdevelopers do not take into account resource shortcomingsof mobile devices in developing mobile application and userswill not consider their devices weaknesses in utilizing variedcomplex applications

B Motivation

Mobile devices have recently gained momentous ground inseveral communities like governmental agencies enterprisessocial service providers (eg insurance Police fire depart-ments) healthcare education and engineering organizations[67] [68] However despite of significant improvement inmobilesrsquo computing capabilities still computing requirementsof mobile users especially enterprise users is not achieved

Several intrinsic deficiencies of mobile devices encumberfeasibility of intense mobile computing and motivate aug-mentation Leveraging augmentation approaches vision ofperforming intense mobile operations and control such asremote surgery on-site engineering and visionary scenariossimilar to the lost child and disaster relief described in [69]will become reality In this part we analyze and taxonomizesmartphonesrsquo deficits that can be alleviated by augmentationFigure 2 depicts our devised taxonomy

1) Processing PowerProcessing deficiencies of mobileclients due to slow processing speed and limited RAM is oneof the major challenges in mobile computing [69] Mobile de-vices are expected to have high processing capabilities similarto computing capabilities of desktop machines for performingcomputing-intensive tasks to enrich user experience Realizingsuch vision requires powerful processor being able to performlarge number of transactions in a short course of time

Large internal memoryRAM to store state stack of allrunning applications and background services is also lackingBeside local memory limitations memory leakage also inten-sifies memory restrains of mobile devices Memory leakageis the state of memory cells being unnecessarily occupied byrunning applications and services or those cells that are notreleased after utilization Garbage-collector-based languageslike Java in Android2 contribute to memory leakage due tofailed or delayed removal of unused objects from the memory[70] Androidrsquos kernel level transactions can also leak memoryin the absence of memory management mechanisms [70][71] Moreover inward deficiency and inefficient design andimplementation of mobile applications can also waste scarcememory of mobile devices Thus in the absence of requiredmemory applications are frequently paused or terminatedby the operating system leading to longer execution timeexcessive resource dissipation and ultimately mobile userexperience degradation

2) Energy ResourcesEnergy is the only non-replenishableresource in mobile devices that demands external resourcesto be replenished [72] [73] Currently energy requirement ofa mobile device is supplied via lithium-ion battery that lastsonly few hours if device is computationally engaged Batterycapacity is increasing at about 5 to 10 a year [74] [75]as battery cells are excessively dense [72] Moreover mo-bile device manufacturers endeavor to attain device lightnesscompactness and handiness which prevent exploitation ofbulky long-lasting batteries User safety is another concern thatconfines manufacturers to produce low capacity batteries [76]While explosion of a battery with few hundreds milliamperescapacity can jeopardize human life [77] explosion of a high-capacity battery can carry catastrophic consequences

Energy harvesting efforts [78]ndash[80] seek to replenish energyfrom renewable resources particularly human movement solarenergy and wireless radiation but these resources are mostlyintermittent and not available on-demand [81] For instancea sitting mobile user at night cannot have any external powersource in the absence of wall power and wireless radiationsMoreover researchers aim at reducing the energy overheadin different aspects of computing including hardware OSapplication and networking interface [82] [83] Effortsare di-rected to develop alternative energy resources such as nuclearbatteries that will likely last months or years [84] Howeversignificant deal of RampD is needed to fulfill ever-increasingenergy requirements of mobile users

Hence in the absence of long-spanning energy on mobiledevices alternative augmentation approaches play a vitalrolein maturing mobile and ubiquitous computing

3) Local Storage Drastic increasing in the number ofapplications and amount of digital contents such as picturessongs movies and home films [85] from one hand and limitedstorage of mobile devices from the other hand decelerateusability of mobile devices While PCs are able to locallystore huge amount of data smartphones are limited to fewgigabytes of space which are mostly occupied by systemfiles user applications and personal data Therefore frequent

2urlhttpwwwandroidcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 6

Fig 2 Taxonomy of Augmentation Motivation Intrinsic andnon-intrinsic mobile challenges motivate augmentation

storing updating and deleting data as well as uninstalling andreinstalling applications due to space limitation cause irksomeimpediments to mobile users [86] Additionally deliveringoffline usability which is one of the most important character-istics of contemporary applications remains an open challengesince mobile devices lack large local storage

4) Visualization CapabilitiesEffective data visualizationon small mobile devicesrsquo screen is a non-trivial task whencurrent screen manufacturing technologies and energy limita-tions do not allow significant size extensions without losingdevice handiness Currently smartphones like HTC One X3

and Samsung Galaxy Note II4 have the biggest screens at47 and 55 inches respectively however they are very smallcompared to PCs and notebooks

Therefore efficient data visualization in small smartphonesrsquoscreen necessitates software-based techniques similar totab-ular pages 3D objects multiple desktops switching betweenlandscape and portrait views (needs accelerometer) and verbalcommunication to virtually increase presentation area Re-cently computing-intensive mobile 3D display technologyispromising to noticeably mitigate the visualization deficitofcontemporary smartphones Glass-free auto-stereoscopicdis-plays [87] can present 3D data by exploiting binocular parallaxto offer a different view for each eye Taking advantagesof current and imminent software-based techniques besidenative tools especially tilting sensors significantly improve themobile visualization capabilities in the near future Howeverthese approaches are computation-intensive processes thatquickly drain battery [87] [88] A feasible alternative solutionto realize software-based content presentation approaches is toaugment smartphonesrsquo computing capabilities

5) Security Privacy and Data SafetyMobile end-usersare concerned about security and privacy of their personaldata banking records and online behaviors [89] The dramaticincrease in cybercrime and security threats within mobiledevices [90] cloud resources [91] and wireless transactionsmakes security and privacy more challenging than ever [92]Moreover performing complex cryptographic algorithms islikely infeasible because of computing deficiencies of mobile

3httpwwwhtccomwwwsmartphoneshtc-one-x4httpwwwsamsungcommyconsumermobile-devicesgalaxy-

notegalaxy-noteGT-N7100RWDXME

devices Securing files using pair of credentials is also lessrealistic in the absence of large keyboard

Data safety is another challenge of mobile devices becauseinformation stored inside the local storage of mobile devicesare susceptible to safety breaches due to high probability ofhardware malfunction physical damage stealing and loss

Amalgam of these problems and deficiencies in mobilecomputing stimulates researchers from academia and industryto exploit novel technologies and approaches to augmentcomputing capabilities of mobile devices which is subject ofthis study

C Mobile Augmentation Types Taxonomy

In this Section we analyze and classify augmentation ap-proaches into two major types of hardware and software Ourdevised taxonomy is depicted in Figure 3 and described asfollowsHardware The hardware approach aims to empower smart-phones by exploiting powerful resources particularly multi-core CPUs with high clock speed [93] large screen and long-lasting battery [84] [94] ARM5 and Samsung6 are major mo-bile processor manufacturers producing multi-core processorssuch as ARM Cortex-597 and Samsung Exynos 5 Octa core8

that perform in higher speed than a single core processors[93] However doubling the CPU clock speed approximatelyoctuples the device energy consumption [66]

Nevertheless augmentation via sophisticated hardware ishindered by several obstacles Firstly generating powerfulprocessor large storage and big screen decrease smartphonehandiness due to additional heat size and weight Secondlyconsidering the fact that utilizing long-lasting battery in smallmobile devices is not feasible with current technologies re-source enlargement contributes toward faster battery drainageand shorter battery life time Thirdly equipping mobile de-vices with high-end hardware noticeably increases their price

5httparmcom6httpsamsungcom7httpwwwarmcomproductsprocessorscortex-a50cortex-a57-

processorphp8httpwwwsamsungcomglobalbusinesssemiconductorminisiteExynos

blog CES 2013 SamsungMobilizes Possibility with Exynos5Octahtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 6: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 6

Fig 2 Taxonomy of Augmentation Motivation Intrinsic andnon-intrinsic mobile challenges motivate augmentation

storing updating and deleting data as well as uninstalling andreinstalling applications due to space limitation cause irksomeimpediments to mobile users [86] Additionally deliveringoffline usability which is one of the most important character-istics of contemporary applications remains an open challengesince mobile devices lack large local storage

4) Visualization CapabilitiesEffective data visualizationon small mobile devicesrsquo screen is a non-trivial task whencurrent screen manufacturing technologies and energy limita-tions do not allow significant size extensions without losingdevice handiness Currently smartphones like HTC One X3

and Samsung Galaxy Note II4 have the biggest screens at47 and 55 inches respectively however they are very smallcompared to PCs and notebooks

Therefore efficient data visualization in small smartphonesrsquoscreen necessitates software-based techniques similar totab-ular pages 3D objects multiple desktops switching betweenlandscape and portrait views (needs accelerometer) and verbalcommunication to virtually increase presentation area Re-cently computing-intensive mobile 3D display technologyispromising to noticeably mitigate the visualization deficitofcontemporary smartphones Glass-free auto-stereoscopicdis-plays [87] can present 3D data by exploiting binocular parallaxto offer a different view for each eye Taking advantagesof current and imminent software-based techniques besidenative tools especially tilting sensors significantly improve themobile visualization capabilities in the near future Howeverthese approaches are computation-intensive processes thatquickly drain battery [87] [88] A feasible alternative solutionto realize software-based content presentation approaches is toaugment smartphonesrsquo computing capabilities

5) Security Privacy and Data SafetyMobile end-usersare concerned about security and privacy of their personaldata banking records and online behaviors [89] The dramaticincrease in cybercrime and security threats within mobiledevices [90] cloud resources [91] and wireless transactionsmakes security and privacy more challenging than ever [92]Moreover performing complex cryptographic algorithms islikely infeasible because of computing deficiencies of mobile

3httpwwwhtccomwwwsmartphoneshtc-one-x4httpwwwsamsungcommyconsumermobile-devicesgalaxy-

notegalaxy-noteGT-N7100RWDXME

devices Securing files using pair of credentials is also lessrealistic in the absence of large keyboard

Data safety is another challenge of mobile devices becauseinformation stored inside the local storage of mobile devicesare susceptible to safety breaches due to high probability ofhardware malfunction physical damage stealing and loss

Amalgam of these problems and deficiencies in mobilecomputing stimulates researchers from academia and industryto exploit novel technologies and approaches to augmentcomputing capabilities of mobile devices which is subject ofthis study

C Mobile Augmentation Types Taxonomy

In this Section we analyze and classify augmentation ap-proaches into two major types of hardware and software Ourdevised taxonomy is depicted in Figure 3 and described asfollowsHardware The hardware approach aims to empower smart-phones by exploiting powerful resources particularly multi-core CPUs with high clock speed [93] large screen and long-lasting battery [84] [94] ARM5 and Samsung6 are major mo-bile processor manufacturers producing multi-core processorssuch as ARM Cortex-597 and Samsung Exynos 5 Octa core8

that perform in higher speed than a single core processors[93] However doubling the CPU clock speed approximatelyoctuples the device energy consumption [66]

Nevertheless augmentation via sophisticated hardware ishindered by several obstacles Firstly generating powerfulprocessor large storage and big screen decrease smartphonehandiness due to additional heat size and weight Secondlyconsidering the fact that utilizing long-lasting battery in smallmobile devices is not feasible with current technologies re-source enlargement contributes toward faster battery drainageand shorter battery life time Thirdly equipping mobile de-vices with high-end hardware noticeably increases their price

5httparmcom6httpsamsungcom7httpwwwarmcomproductsprocessorscortex-a50cortex-a57-

processorphp8httpwwwsamsungcomglobalbusinesssemiconductorminisiteExynos

blog CES 2013 SamsungMobilizes Possibility with Exynos5Octahtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 7: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 7

Fig 3 Taxonomy of Mobile Augmentation Types

compare to the stationary machines Unlike PCs smartphonersquoshardware is not upgradable hence a new device shouldbe possessed in case of technology advancement Thereforein the absence of futuristic engineering technologies thehardware-based augmentation process is slow and expensivethat necessitates alternative augmentation approaches toen-hance computing capabilities of mobile devices without drasticownership price hikeSoftwareSoftware-oriented mobile augmentation approachesare classified into five groups and will be explained later inthis part Resources that are used in major software-orientedapproaches are classified into two groups namely traditionaland cloud-based Their major differences lie on resource pro-visioning and access strategies service security and deliverymodels and resource characteristics In traditional approachesresearchers leverage centralized resources of distant traditionalservers or free nearby surrogates Several problems suchas resource availability elasticity and security of traditionalapproaches hinder their success For instance surrogatescanterminate their services anytime without considering theircurrent load and can violate user security and privacy bychanging execution sequence or altering raw and processeddata

To alleviate the problems of traditional servers researchersin recent efforts [25] [27] [29] [31] [33]ndash[35] [41] [43][44] [95] exploit highly available elastic secure cloudin-frastructure ldquoCloud is a type of parallel and distributedsystem consisting of a collection of interconnected and vir-tualized computers dynamically provisioned and presentedasone or more unified computing resources based on service-level agreements established through negotiation betweentheservice provider and consumersrdquo [20]

While utilizing cloud resources users pay for the amountand duration they utilize various resources (eg CPU mem-ory and bandwidth) based on an agreed SLA In the SLA theamount and quality of required resources such as processorRAM and storage are specified and user is billed accordinglyService delivery failure will be compensated by the vendorLucrative financial benefits of cloud services encourage cloudproviders to compete in delivering high service availabilityreliability security and robustness to increase their marketshare Hence the augmentation performance is less affected

by resource unreliability and interruptionMoreover cloud infrastructures are available to end-users

via Virtual Machine(VM)9 to increase resource utilizationratio and enhance overall security and privacy Virtualizationtechnology aims to enable resource sharing in an isolatedenvironment called VM It realizes execution of multipleoperating systems on a single machine and enables sharingof large resources among multiple end-users Users can onlyaccess to infrastructures allocated to their VMs and cannotaccess prohibited resources and contents

Table III summarizes the comparison results of traditionaland cloud-based resources and advocates differentiationsbe-tween the conventional servers and clouds High computingpower elasticity mobility support low utilization overheadand security are some of the significant advantages of cloudresources compare to the surrogates that advocate the latestparadigm shift in mobile augmentation

Software augmentation techniques are classified as remoteexecution (offloading or cyber foraging) [5]ndash[8] [10]ndash[13][16]ndash[18] [25] [29] [30] [33]ndash[35] [41] [43] [44] [96]remote storage [97] Multi-tier programming [36] [45] [46]live cloud-streaming [98] resource-aware computing [99][100] and fidelity adaptation [101] and explained as follows

bull Remote executionAs explained in II-Athe resource-hungry components of mobile applications mdashin whole orpart mdashare migrated to the resource-rich computing device(s)that are willing to share their resources with mobile devicesRapid development of heterogeneous mobile devices obligesadaptive offloading approaches able to enhance capabilitiesof wide range of mobile devices in dynamic environmentwith least processing overhead and latency The efficiency ofoffloading approaches highly depends on what component(s)can be partitioned When partitioning takes place Where toexecute the component(s) And how to communicate with theremote server [102] Offloading approaches perform varied-time analysis to answer these questions which are classifiedinto three groups and explained as below

Design Time AnalysisIn this method the applicationrsquoscomplexity is analyzed at design time to determine the answerof four above questions Application developer or a middle-ware specifies the resource-intensive components of mobile

9httpwwwvmwarecomvirtualizationwhat-is-virtualizationhtml

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 8: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 8

TABLE IIICOMPARISONBETWEEN TRADITIONAL AND CLOUD-BASED COMPUTING RESOURCE

Features Traditional Cloud-BasedComputation Power Low HighElasticity Low HighUser Experience Low HighReliability Low HighAvailability Intermittent On-demandClient Mobility Limited UnlimitedMulti-tenancy Not available AvailableServing Incentive Not provisioned ProvisionedUtilization Cost Free Pay-As-You-UseUtilization Overhead High LowManagement Decentralized DeCentralizedBack-end Connectivity Wired Wired amp WirelessCommunication Latency Low VariedComputation Latency High LowSecurity Low HighData Safety Low High

application that can be offloaded to the remote server and labelthem as remote component(s) Programmers decide how topartition application and adapt its performance to the dynamicmobile environment which is a non-trivial task mainly dueto the lack of knowledge about the execution environmentPerforming such action needs excessive programming skilland knowledge of computation offloading Design time ap-proaches [8] [10] [12] notably save native resources ofmobile device by reducing the processing and monitoringoverheads However partitioning prior to the execution isnotalways optimal and cannot accurately adapt performance indiverse execution environments and also imposes extra effortson the application developer or middleware for deciding onpartitioning Hence design time partitioning approachesarelikely become obsolete

Runtime AnalysisRuntime or dynamic partitioning referredto methods such as [25] [103] that aims to answer fourquestions at runtime They identify and partition the resource-hungry parts of the application specify how and where toexecute the partitioned components [102] [104] and de-termine how to communicate with the server In dynamicmethods resource requirement of the application is analyzedand available resources are detected to decide if the appli-cation requires remote resources Upon decision making thesystem performs offloading Further monitoring is necessaryto gather knowledge of available remote resources to maintainexecution history Although these approaches provide dynamicand flexible solutions large amount of resources are wastedat runtime that prolongs application execution and decreaseenergy efficiency

Hybrid AnalysisThe ultimate aim of hybrid approaches[105] is to increase performance and efficiency of augmen-tation methods Deciding on how to perform the offloadingmainly depends on the native resources remote resources andavailable network bandwidth In [105] prior to the applicationexecution the system decides based on four options namelyi)

no action ii) dynamic iii) static and iv) profile only whetherto offload or not and in case of offloading specify what typeof partitioning should take place The profile only option issimilar to the no action but the systems collect executioninformation to maintain execution history for future purpose

bull Remote StorageRemote storage is the process of ex-panding storage capability of mobile devices using remotestorage resources It enables maintaining applications anddata outside the mobile devices and provides remote accessto them In early efforts researchers in [97] utilize iSCSI(Internet Small Computer System Interface) [106] mdashas awell-established protocol for remote storage mdashto access theserverrsquos IO resources via mobile clients over the TCPIPnetwork to store backup and mirror data [107] Howeverthe throughput of iSCSI is highly affected by the mobile-server distance Using iSCSI is also difficult for handlinglarge files such as multimedia and database files Moreoverdue to message passing in wireless medium through TCPIPthe security and processing overhead (eg cryptography anddata compression) are further challenges To alleviate thesechallenges several researches as MiSC [108] UbiqStor [109][110] and Intermediate Target [111] are proposed towardsrealizing remote storage on mobile devices However due toscalability availability performance and efficiency issues oftraditional servers power of remote storage could not fullyunleash using traditional servers

Several proposals and data storage services in academiaand industry aim to expand mobile storage by exploitingcloud computing especially Jupiter [31] SmartBox [112]Amazon S310 Mozy11 Google Docs12 and DropBox13 Forinstance Jupiter expands smartphone storage and assists end-users in organizing large applications and data Jupiter lever-

10httpawsamazoncoms311httpmozycom12httpsdocsgooglecom13httpswwwdropboxcom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 9: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 9

ages cloud infrastructures to store big data of mobile usersHeavy applications are executed inside the cloudrsquos VM ofsmartphones and results are forwarded to the physical deviceafter execution Amazon S3 is a general purpose storageoffering simple operations to store and retrieve cloud datawhile Mozy provides data backup facilities with main focuson enhancing cloud data safety against natural disasters

bull Resource-Aware ComputingIn resource-aware comput-ing efforts especially [99] [100] [113]ndash[115] resource re-quirements of mobile applications are diminished utilizingthe application-level resource management methods (usingapplication management software such as compiler and OS)and lightweight protocols Resource conservation is performedvia efficient selection of available execution approachesand technologies [114] Any mobile application consists ofapplication-level resource management method is consideredas a resource-aware application For instance in [100] authorspropose an energy-friendly scheme for content-based imageretrieval applications using three offloading options namelyi) local extraction-remote search ii) remote extraction-remotesearch and iii) remote extraction-local search The authorsconsider available bandwidth image database size and num-ber of user queries to opt any of three offloading optionsfor saving energy In a high bandwidth network with limitedqueries the third option is beneficial the system uploads allun-indexed images to the remote server and receives the resultsto be loaded into the memory Then all search queries areexecuted locally

Similarly applications can decide whether to choose 2G or3G in telephony and FTP Using 2G network for telephony and3G for FTP can noticeably reduce resource requirements ofthe mobile applications according to the power consumptionpatterns presented in [116] 2G network technology consumesless energy for establishing a telephony communication while3G is more energy-friendly for file transfer transactions

bull Fidelity adaptationFidelity adaptation is an alternativesolution to augment mobile devices in the absence of remoteresources and online connectivity In this method local re-sources are conserved by decreasing quality of applicationexecution which is unlikely desirable to end-users As awell-known fidelity adaptation approach we can refer tothe YouTube14 Users in YouTube can adjust the streamingquality based on available bandwidth To achieve optimizedperformance researchers [78] [117] leverage composition ofcyber foraging and fidelity adaptation

bull Multi-tier Programming Developing distributed multi-tier mobile applications leveraging remote infrastructures isanother technique employed in efforts such as [36] [45] [46][118] to reduce resource requirements of mobile applicationsThe main idea in this type of mobile applications is toreduce the client-side computing workload and develop theapplications with less native resource requirements Certainlythe computationally intensive components of the applicationsare executed outside the device whereas the interactive (userinterface) and native codes (eg accessing to the devicecamera) remain inside the device for execution

14httpyoutubecom

Multi-tier applications are lightweight aiming to consumethe least possible local resources by utilizing remote compo-nents and services whereas native applications are monolithicapplications often require runtime migration for executionTherefore monitoring time and communication overhead ofmulti-tier applications are shrunk leading to explicit resourcesaving and user experience enhancement

bull Live Cloud StreamingIn recent efforts to harness cloud re-sources researchers fromOnlive15 andGaikai16 among otherorganizations introduce new approach to augment computingcapabilities of mobile devices entitled live cloud streaming[98] In live cloud streaming approaches mobile device actsas a dump client able to interact with server using a browseror application GUI In live cloud streaming applications entireprocessing take place in the cloud and results are streamingto the mobile devices However usability of cloud-streamingis hindered by latency network bandwidth portability andnetwork traffic cost

Functionality of cloud-streaming applications absolutely de-pends on the network availability and the Internet Transferringmobile-user input to the server is another critical factor thatrequires considerable attention under wireless Internet connec-tion Moreover since majority of mobile network providersdeploy lsquopay-as-you-usersquo data plans the large data traffic ofcloud-streaming services imposes high communication coston users Yet congestion handling remains an open issue atpeak hours Entirely relying on cloud-streaming infrastructuresand avoiding smartphones resourcesrsquo utilization impact onapplication responsiveness and levy extravagant ownershipmaintenance power and networking expenses to the cloud-streaming service providers which is not a green computingapproach

III I MPACTS OFCMA ON MOBILE COMPUTING

This Section discusses the advantages and disadvantagesof performing a CMA process on mobile computing thatare summarized in Table IV We aim to demonstrate howCMA approaches mitigate deficiencies of mobile computingexplained in Section II-B In this Section the terms lsquocloudresourcesrsquo and lsquocloud infrastructuresrsquo refer to any type ofcloud-based resources and infrastructures discussed in SectionV

A Advantages

In this part eight major benefits of utilizing cloud resourcesin mobile augmentation processes are introduced

1) Empowered ProcessingEmpowering processing is thestate of virtually increased transaction execution per secondand extended main memory leveraging CMA approaches Incomputing-intensive mobile applications either the hostingdevice does not have enough processing power and memoryor cannot provide required energy A common solution is tooffload the application mdashin whole or part mdashto a reliablepowerful resource with least energy and time cost In compu-tation offloading the complex CPU- and memory-intensive

15httponlivecom16httpgaikaicom

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 10: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 10

components of a standalone application are migrated to thecloud Consequently the mobile devices can virtually performand actually deliver the results of heavy transactions beyondtheir native capabilities

Although surrogates in traditional augmentation approaches[8]ndash[10] [12] could increase computing capabilities of mobiledevices excessive overhead of arbitrary service interruptionand denial could shadow augmentation benefits [19] [119]Cloud resources guarantee highest possible resource availabil-ity and reliability

Leveraging CMA approaches application developers buildmobile application with no consideration on available nativeresources of mobile devices and mobile users dismiss theirdevicesrsquo inabilities Hence computing- and memory-intensivemobile applications like content-based image retrieval appli-cations (enable mobile users to retrieve an image from thedatabase) can be executed on smartphones without excessefforts

However a flexible and generic CMA approach that canenhance plethora of mobile devices with least configurationprocessing overhead and latency is a vital need in excessivelydiverse mobile computing domain Such diversity is mainlydue to the rapid development of smartphones and Tabletsand sharp rise in their hardware platform API feature andnetwork heterogeneity [120] in the absence of early standard-ization

2) Prolonged Battery Long-lasting battery can be con-sidered as one the most significant achievements of CMAapproaches for large number of mobile users Smartphonemanufacturers have already utilized high speed multi-coreARM processors (eg Cortex-A57 Processor17) being able toperform daily computing needs of mobile end-users How-ever such giant processing entities consume large energyand quickly drain the battery that irks end-users CMA so-lutions can noticeably save energy [95] by migrating heavyand energy-intensive computing to the cloud for executionAlthough energy efficiency is one of the most importantchallenges of current CMA systems several efforts such as[53]ndash[55] [121] [122] are endeavoring to comprehend theenergy implications of exploiting cloud-based resources frommobile devices and shrinking their energy overhead

In traditional cyber foraging or surrogate computing ap-proaches energy is saved by computation offloading butseveral issues such as lack of mobility support and resourceelasticity can neutralize the benefits of energy-hungry taskoffloading

3) Expanded StorageInfinite cloud storage accessiblefrom smartphones enables users to utilize large number ofapplications and digital data on device Hence they are notobliged to frequently install and remove popular applicationsand data due to the space limit Online connectivity is essentialto access cloud storage In such online storage systems dataare manually or automatically updated to the online storagefor maintaining the consistency of the online storage systemStoring applications in cloud storage provides the opportu-nity to update the code without consuming any mobile IO

17httpwwwarmcomproductsprocessorscortex-a50cortex-a57-processorphp

transactions which enhances user experience and improves thesmartphonesrsquo energy efficiency mdashbecause IO transactions areenergy-hungry tasks

4) Increased Data SafetyCMA efforts can bring thebenefit of data safety to the mobile users Naturally storeddata on mobile devices are susceptible to loss robberyphysical damage and device malfunction Storing sensitiveand personal data such as online banking information onlinecredentials and customer related information on such a riskystorage significantly degrades the quality of user experienceand hinders usability of mobile devices Due to the scarcecomputing resources especially energy in mobile devicesperforming complex and secure encryption provisions is notfeasible Hence by storing data in a reliable cloud storage[112] [123] users ensure data availability and safety regard-less of time place and unforeseen mishaps Threats such asdevice robbery or physical damage to the mobile devices willeffect on the tangible value of the device rather than intangiblevalue of the data

5) Ubiquitous Data Access and Content SharingCloudinfrastructures play a vital role in enhancing data accessquality Storing data in cloud resources enables mobile usersto access their digital contents anytime anywhere from anydevice Hence the impact of temporal geographical andphysical differences is noticeably decreased that enriches userexperience

Moreover cloud storages facilitate data sharing and contri-bution among authorized users Every file and folder in cloudusually has a protected unique access link that can be obtainedby the owner to share them among legitimate users Networktraffic is hence shrunk because data is accumulated in a centralserver accessible to unlimited users from various machinesCloud can significantly enhance data transfer among differentmobile devices One of the most irksome userrsquos impedimentsis to transfer data from current mobile device to a new handsetApart from its temporal cost porting data from one device toanother especially to a heterogeneous device is a risky practicethat puts data is in the risk of corruption and loss of integrityStored data on Cloud remain safe and can be synchronized toany number of mobile devices with minimum risk Howevera reliable data access control mechanism is required to adjustuser permissions

6) Protected Offloaded ContentCloud storage solutionsaim to protect remote codes and data while ensure userrsquosprivacy This is one of the most important gains of replacingsurrogates with cloud resources Cloud servers deploy virtu-alization technology to isolate the guest environment fromother guests and also from their permanent software stackMoreover cloud vendors deploy strict security and privacypolicies to not only ensure confidentiality of user contentbut also to protect their properties and business Implementinternal security provisions particularly the state-of-the-art bio-metric security systems to protect their physical infrastructuresand avoid unauthorized access Employing complex contentencryption frequent patching and continuous virus signatureupdate inside the company premise or seeking technical ser-vices from a trusted third party [124] are other examples ofsecurity provisions undertaken in cloud to further protectcloud

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 11: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 11

TABLE IVIMPACT OF CMA A PPROACHES INMOBILE COMPUTING

Advantages DisadvantagesEmpowered Processing Dependency to High Performance Networking Infrastructure

Prolonged Battery Excessive Communication Overhead and TrafficExpanded Storage Unauthorized Access to offloaded Data

Increased Data Safety Application Development ComplexityUbiquitous Data Access Paid Infrastructures

and Content SharingProtected Offloaded Contents Inconsistent Cloud Policies and Restrictions

Enriched User Interface Service Negotiation and ControlEnhanced Application Generation Nil

storage7) Enriched User InterfaceAs described in II-B4 vi-

sualization shortcomings of mobile devices diminish userexperience and hinder smartphonesrsquo usability However cloudresources can be exploited to perform intensive 2D or 3Dscreen rendering The final screen image can be prepared basedon the smartphone screen size and streamed to the deviceConsequently screen adaptation also is achieved when cloudside processing engine automatically alter the presentationtechnique to match screen image with the device screen size

8) Enhanced Application GenerationCloud resources andcloud-based application development frameworks similar tomicroCloud and CMH facilitate application generations in het-erogeneous mobile environment Once a cloud component isbuilt it can be utilized to develop various distributed mobileapplications for large number of dissimilar mobile devices Inthe presence of cloud components application programmerneeds to develop native mobile components and integratethem with relevant prefabricated cloud components to developa complex application When a mobile-cloud application isdeveloped for Android device by slightly changing nativecomponents the application is transited to new OS like iOS18

and Symbian19 which significantly save time and money

B Disadvantages

Despite of many advantageous aspects of cloud servicestheir success is hindered by several drawbacks and shortcom-ings that are discussed as follows

1) Dependency to High Performance Networking Infras-tructure CMA approaches demand converged wired andwireless networking infrastructures and technologies to fulfillintersystem communication requirements In wireless domainCMAs need high performance robust reliable high band-width wireless communication to realize the vision of com-puting anywhere anytime from any-device In wired commu-nication fast reliable communications ground is essential tofacilitate live migration of heavy data and computations toaregional cloud-based resources near the mobile users Effortssuch as next generation wireless networks [125] and the openmobile infrastructure [126] with Open Wireless Architecture

18httpwwwapplecomios19httplicensingsymbianorg

(OWA) by Sieneon [127] contribute toward enhancing thenetworking infrastructuresrsquo performance in MCC

2) Excessive Communication Overhead and TrafficMobiledata traffic is significantly growing by ever-increasing mo-bile user demands for exploiting cloud-based computationalresources Data storageretrieval application offloading andlive VM migration are example of CMA operations thatdrastically increase traffic leading to excessive congestionand packet loss Thus managing such overwhelming trafficand congestion via wireless medium becomes challengingespecially when offloading mobile data are distributed amonghelping nodes to commute tofrom the cloud Consequentlyapplication functionality and performance decrease leading touser experience degradation

3) Unauthorized Access to Offloaded DataSince cloudclients have no control over their remote data users contentsare in risk of being accessed and altered by unauthorizedparties Migrating sensitive codes as well as financial andenterprise data to publicly accessible cloud resources decreasesusers privacy especially enterprise users Moreover storingbusiness data in the cloud is likely increasing the chanceof leakage to the competitor firm Hence users especiallyenterprise users hesitate to leverage cloud services to augmenttheir smartphones

4) Application Development ComplexityThe excessivecomplexity created by the heterogeneous cloud environmentincreases environmentrsquos dynamism and complicates mobileapplication development Mobile application developers arerequired to acquire extensive knowledge of cloud platforms(ie cloud OSs programming languages and data structures)to integrate cloud infrastructures to the plethora of mobiledevices Understanding and alleviating such complexity im-pose temporal and financial costs on application developersand decrease success of CMA-based mobile applications

5) Paid InfrastructuresUnlike the free surrogate resourcesutilizing cloud infrastructure levies financial charges totheend-users Mobile users pay for consumed infrastructuresaccording to the SLA negotiated with cloud vendor In certainscenarios users prefer local execution or application termina-tion because of monetary cloud infrastructures cost Howeveruser payment is an incentive for cloud vendors to maintaintheir services and deliver reliable robust and secure servicesto the mobile users

In addition cloud vendors often charge mobile users twice

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 12: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 12

Fig 4 Taxonomy of Cloud-based Computing Resources

once for offloading contents to the cloud and once again whenusers decide to transfer their cloud data to another cloudvendors to utilize more appropriate service (eg monetary andQoS (Quality of Service) aspects)

6) Inconsistent Cloud Policies and RestrictionsOne of thechallenges in utilizing cloud resources for augmenting mobiledevices is the possibility of changes in policies and restrictionsimposed by the cloud vendors Cloud service providers applycertain policies to restrain service quality to a desired level byapplying specific limitations via their intermediate applicationslike Google App Engine bulk loader20 Services are controlledand balanced while accurate bills will be provided based onutilized resources

Also service provisioning controlling balancing andbilling are often matched with the requirements of desktopclients rather than mobile users [128] Considering the greatdifferences in wired and wireless communications disregard-ing mobility and resource limitations of mobile users indesign and maintenance of cloud can significantly impact onfeasibility of CMA approaches Hence it is essential to amendrestriction rules and policies to meet MCC users requirementsand realize intense mobile computing on the go

7) Service Negotiation and ControlWhile cloud users arerequired to negotiate and comply with the cloud terms andconditions for a certain period of time often cloud agreementsare nonnegotiable and policies might change over the timeMoreover there is no control over the cloud performance andcommitments in the absence of a controlling authority or atrusted third party Hence CMA services are always volatileto the service quality of cloud vendors

IV TAXONOMY OF CLOUD-BASED COMPUTING

RESOURCES

Researchers [24]ndash[27] [27]ndash[43] [45]ndash[49] aim to obtainuser requirements and preferences by exploiting varied types

20httpsdevelopersgooglecomappenginedocspythontoolsuploadingdata

of cloud-based resources to augment computing capabilitiesof resource-constraint smartphones Based on the distanceand mobility traits of such varied cloud-based computingresources we classify them into four groups namely distantimmobile clouds proximate immobile computing entitiesproximate mobile computing entities and hybrid that aretaxonomized in Figure 4 and explained as follows Table Vrepresents the comparison results of these cloud-based com-puting resources This Table can be utilized as a guideline forappropriate selection of cloud-based infrastructures in futureCMA researches

A Distant Immobile Clouds

Public and private clouds comprised of large number ofstationary servers located in vendors or enterprises premisesare classified in this category They are highly availablescalable and elastic resources that are often located far fromthe mobile nodes accessible via the Internet Although publiccloud resources are likely more secure compared to the othertypes of resources due to complex security provisions andon-premise infrastructures [129]ndash[132] they are vulnerableto security attacks and breaches like Amazon EC2 crash[92] and Microsoft Azure security glitch [133] Accessingcloud resources especially public clouds often carries therisk of communicating through the risky channel of InternetHowever giant clouds are endeavoring to maintain security-for more market share- and could establish high reputation-based trust by providing long-term services to the users

Additionally the performance and efficacy of these ap-proaches are affected by long WAN latency due to the longdistance between mobile client and stationary cloud data cen-ters One potential approach to shorten the distance betweenmobile device and cloud is to migrate the remote code anddata to the computing resources near to the mobile devicevia live migration of the VM from the cloud [134] Howeverlive migration of VM is a non-trivial task that requires greatdeal of research and development particularly in networkingenvironment due to several issues such as large VM size

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 13: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 13

TABLE VCOMPARISON OFCLOUD-BASED SERVERS

Distant clouds Proximate immobile Proximate mobile Hybridcomputing entities computing entities

Architecture DistributedOwnership Service provider Public Individual Hybrid

Environment Vendor Premise Business Center Urban Area HybridAvailability High Medium Medium HighScalability High Medium Medium High

Sensing Capabilities Medium Low High HighUtilization Cost Pay-As-You-Use

Computing Heterogeneity High Medium High HighComputing Flexibility High Medium High High

Power Efficiency High Medium Medium HighExecution Performance High Medium Medium High

Security and Trust High Moderate Low HighUtilization Rate High

Execution Platform VM VM PhysicalVM PhysicalVMResource Intensity High Moderate Moderate Rich

Complexity Low Moderate Moderate HighCommunication Technology 3GWiFi WiFi WiFi 3GWiFi

Communication Latency High Low Low ModerateExecution Latency Low Medium Medium Low

Maintenance Complexity Low Medium Medium High

hard-to-predict user mobility pattern and limited intermittentwireless bandwidth

Resource utilization is enhanced in clouds due to the virtual-ization technology deployment Several VMs can be executedon a single host to increase the utilization efficiency of theclouds while each computation task runs on a single isolatedVM loaded on a physical machine However VM securityattacks such as VM hopping and VM escape [135] can violatethe code and data security VM hopping is a virtualizationthreat to exploit a VM as a client and attack other VM(s) onthe same host VM escape is the state of compromising thesecurity of the hypervisor and control all the VMs

B Proximate Immobile Computing Entities

The second type of cloud-based computing resources in-volves stationary computers located in the public places nearthe mobile nodes The number of computers in public placessuch as shopping malls cinema halls airports and coffeeshops is rapidly increasing These machines are hardly per-forming tense computational tasks and are mostly playing mu-sic showing advertisement or performing lightweight appli-cations Moreover they are connected to the power socket andwired Internet Therefore it is feasible to leverage such abun-dant resources in vicinity and perform extensive computationon behalf of resource-constraint mobile devices It can alsoreduce latency and wireless network traffic while increasesresource utilization toward green computing Another group ofproximate immobile computers are Mobile Network Operators(MNO) and their authorized dealers scattered in urban andrural areas private clouds and public computing kiosk [136]that can be exploited in smartphone augmentation

However protecting security and privacy of mobile user andcomputer owner hinder utilization of such nearby resourcesSeveral shortcomings such as insufficient on-premise securityinfrastructure lake of tight security mechanisms and inef-ficient update and maintenance procedures inhibit utilizingsuch resources (except MNOs) for CMA approaches Ownersof these resources may attack mobile users and access theirprivate data on the mobile devices or falsify offloading resultsAlso malicious users may leverage these resources as anattacking point to violate mobile usersrsquo security and privacyOn the other hand security and privacy of resource ownersare also susceptible to violation Owners of computer devicesparticipating in resource sharing require robust mechanismsto protect and isolate the guest code and data from theirhost applications and data Virtualization aims to realizesuchisolation mechanism but issues such as VM hopping and VMescape require to be addressed before its successful adoption[135] Among all proximate immobile resources MNOs maybe considered unique in terms of security and privacy featuresMNOs in general have been serving mobile users for longtime and could establish high degree of trust among mobileusers It is feasible to assume that MNOrsquos certified dealers alsocan inherit MNOrsquos trust if central management and monitoringprocess is undertaken by MNOs [46]

C Proximate Mobile Computing Entities

In this category of cloud-based infrastructures variousmobile devices particularly smartphones Tablets notebookswearable computers and handheld computing devices playthe role of servers based on cloud computing principles Themain benefit of utilizing nearby mobile resources is their

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 14: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 14

Fig 5 The Hybrid Cloud Concept for MCC

proximity to the mobile clients Also hardware and platformheterogeneity [51] between mobile servers and clients can bemitigated because both sides are mainly ARM-based deviceswith mobile OSs Moreover contemporary smartphones areable to provide value added context- and social-aware ser-vices [137] [138] that contribute to the context-awarenessof mobile applications However mobile devicesrsquo resourcesare limited and they are unable to perform intensive context-computing [139] Realizing distributed computing on clusterof nearby mobile devices requires several issues particularlyapplication architectures resource scheduling and mobility tobe addressed

Moreover security and privacy of mobile devices as aservice provider is a critical concern in CMA Mobile devicesare intrinsically susceptible to loss and robbery and their con-straint resources inhibit exploiting robust security mechanismsinside the device Furthermore with ever-increasing popularityof mobile Apps (ie mobile applications) in online App storessuch as Google Play21 and Samsung APPs22 [140] numberof mobile security threats are rising sharply and malware-contaminated Apps are becoming serious threats to the mobileusers [141] Several security threats have been identified in anexperiment of Android mobile applications with the potentialto violate the security of mobile users [142] Risk of suchcontaminated codes can likely be transferred to the non-contaminated mobile devices by utilizing their computationresources and request for results of a remote computationHence establishing trust between mobile devices and end-users becomes a challenging task

21httpsplaygooglecomstore22httpwwwsamsungappscom

D Hybrid (Converged Proximate and Distant Computing En-tities)

Hybrid infrastructures as depicted in Figure 5 are comprisedof various proximate and distant computing nodes eithermobile or immobile The main idea behind building hybridresources is to employ heterogeneous computing resources tocreate a balance between user requirements (mainly latencyand computation power) and available options [143] Thelatency sensitive codes are offloaded to the nearest computingdevice(s) whereas the most intensive and least latency sensitivetasks are migrated to the furthest resources Perhaps theutilization costs of nearby resources are more than the remoteservers

Beneficial characteristics of hybrid resources summarizedin Table V advocates their usefulness in maximizing the aug-mentation benefits However deployment management andresource scheduling processes in dynamic mobile environmentare non-trivial tasks Developing an autonomic managementsystem similar to CometCloud [144] in cloud computing andMAPCloud [143] in MCC to automatically manage optimizeand adapt hybrid infrastructures in the cloud-mobile applica-tions can significantly improve the quality of hybrid CMAapproaches

Hybrid cloud infrastructures can deliver enhanced securityand privacy features to the CMA approaches and increase theQoS Hybrid clouds are comprised of resources with variedsecurity privacy and trust features which can be efficientlyutilized by CMA and mobile users as a trade-off For instancesecurity sensitive computations can perform a security-latencytrade-off and execute computation inside a secure distantcloud

V THE STATE-OF-THE-ART CMA A PPROACHESTAXONOMY

Cloud-based Mobile Augmentation (CMA) is the-state-of-the-art mobile augmentation model that leverages cloud com-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 15: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 15

Fig 6 Taxonomy of State-of-the-art CMA Models

puting technologies and principles to increase enhance andoptimize computing capabilities of mobile devices by exe-cuting resource-intensive mobile application componentsinthe resource-rich cloud-based resourcesAccording to ourresource classifications in previous Section we analyze andtaxonomize the state-of-the-art CMA approaches into fourmodels namely distant fixed proximate fixed proximatemobile and hybrid which are depicted in Figure 6 Foreach model we describe few CMA efforts and tabulate thecomparison results in Figure 10

A Distant Fixed

Majority of CMA approaches [25] [27] [29] [31] [33]ndash[35] [41] [43] [44] [54] [145] leverage fixed cloud in-frastructures in distance due to its straightforward approachUtilizing stationary cloud eliminates several managementcom-plexities (eg resource discovery and scheduling for mobilecloud-based servers) and alleviates reliability and securityconcerns [18] Works in this class of CMA systems aimat reducing the complexity and overhead of utilizing distantcloud For instance in [54] authors propose an energy-efficientoffline job scheduling model based on makespan minimizationmodel to enhance energy efficiency of distant fixed CMAsystems Their main notion is to separate the data transmissionfrom the job execution During their work authors provideseveral optimization solutions aiming to reduce the energyconsumption of the device during the offloading processHowever for the sake of simplicity the authors study theenergy consumption of tasks in offline mode only which doesnot consider runtime dynamism of MCC

Exploiting cloud resources is feasible in several real sce-narios such as live cloud streaming [98] enterprise appli-

cations (eg Customer Relation Management (CRM) andenterprise resource planning [146]) and Social NetworkingCloud streaming mechanism has already described in II-C asan example of utilizing distance fixed resources In [146]researchers leverage cloud resources in developing a CRMapplication to enhance efficiency of sale representatives for apharmaceutical company The representative meets the physi-cian in medical centers to promote drugs present samples andpromotions material and he records all sale results and detailsthrough the mobile application The huge database of thecompany is stored inside the cloud and the sale representativecan request to process get or update data in database withoutstoring data locally

We describe some of the distant fixed CMA approaches thatutilize distant fixed cloud resources for mobile augmentationas follows The terms immobile fixed and stationary areinterchangeably used with the same notion

bull CloneCloud CloneCloud [34] is a cloud-based fine-grained thread-level application partitioner and execu-tion runtime that clones entire mobile platform into thecloud VM and runs the mobile application inside the VMwithout performing any change in the application codeThe CloneCloud enables local execution of remainingmobile application when remote server is running theintensive components unless local execution tries to ac-cessing the shared memory state Cloud resources in thiseffort simulate distributed execution of a monolithic ap-plication in a resourceful environment without engagingapplication developer into the distributed application pro-gramming domain CloneCloud can significantly reducethe overall execution time using thread-level migrationWhen the local execution reaches the intensive compo-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 16: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 16

nent(s) the CloneCloud system offloads the component(s)to the cloud and continues local execution until theapplication fetches data from the migrated state The localexecution is paused until the results are returned andintegrated to the local applicationHowever the communication overhead of transferring theclone of mobile platform application and memory stateand frequent synchronization of the shared data betweenthe mobile and cloud can shrink the power of cloudSuch overhead becomes more intense in case of heavydata- and communication-intensive and tightly coupledmobile applications where an alternative execution ofresource-intensive and lightweight components existsFrequent code and data encapsulation and migration andmobile-cloud data synchronization excessively increasethe communication traffic and impact on execution timeand energy efficiency of the offloading

bull Elastic ApplicationElastic application model [33] is aCMA proposal leverages distant fixed cloud data cen-ter for executing resource-intensive components of themobile application Authors in this model partition amobile application into several small components calledweblet Weblets are created with least dependency to eachother to increase system robustness while decrease thecommunication overhead and latency The weblet execu-tion is dynamically configured to either perform locallyor remotely based on the webletrsquos resource intensityexecution environment quality and offloading objectivesThe distinctive attribute of this proposal is that applicationexecution can be distributed among more than one ma-chine and cooperative results can be pushed back to thedevice To achieve such goal multiple elasticity patternsnamely replication splitter and aggregator are defined Inreplication pattern multiple replicas of a single interfaceare executed on multiple machines inside the cloudHence failure in one replica will not compromise thesystem performance In splitter pattern the interface andimplementation are separated so that several weblets withvaried implementations can share a single interface Inaggregator the results of multiple weblets are aggregatedand pushed to the device for optimized accuracy andefficacyThe authors endeavor to specify the execution configu-rations (specifying where to run the weblets) at runtimeto match the requirements of the applications and usersTo enhance the overall execution performance and enrichuser experience the system is able to run the weblets bothlocally and remotely A weblet can be executed remotelyin a low-end device while the same can be executedlocally on a high-end deviceElastic application model pays more attention to theuser preferences by enabling different running modes ofa single application (eg high speed low cost offlinemode) However it engages application developers todetermine weblets organization based on the functional-ity resource requirements and data dependency But thecharacteristics of the weblets are mainly inherited fromthe well-known web services to decrease the programmer

burdensbull Virtual Execution Environment(VEE)Hung et al [28]

propose a cloud-based execution framework to offloadand execute the intensive Android mobile applicationsinside the distant cloudrsquos virtual execution environmentThe quality and accuracy of execution environment ishighly influenced by the comprehensiveness and accuracyof emulated platform This method uses a software agentin both mobile and cloud sides to facilitate the overallsystem management The agent in mobile device initiatesVM creation and clones the entire application (even na-tive codes and UI components) and partial datamemorystate from device to the cloud Unlike CloneCloud VEEaims to reduce latency by migrating the segment of datastack explicitly created and owned by the application tothe VM instead of copying the entire memory cloningthe entire memory state especially for heavy applicationssignificantly increases latency and trafficDuring remote execution the system frequently synchro-nizes the changes between device and cloud to keepboth copies updated In order to increase the quality andefficiency of remote execution in virtual environment andavoid data input loss at application suspension stagesthe system stores input events (reading a file capturing aface storing a voice) exploiting a recordreplay schemeand pseudo checkpoint methods However these methodsengage application developers to separate the applicationstate into two states namely global and local and tospecify the global data structures The global state con-tains the program domain and application flow whereasthe local state contains local data structures required bya method Programmer usually needs to identify globalstate when the application is paused Once the applicationis suspended the global state will be loaded to avoid re-execution and the latest Android checkpoint is appliedto the system to reflect all the changes made from thelast checkpoint However all changes especially userinput might be lost from the last checkpoint To recordthe changes after the last checkpoint the recordreplaymechanism is deployed by creating a pseudo checkpointTo create a pseudo checkpoint the application notifiesthe local agent to identify the input events and record re-quired information Upon the application resumption thepseudo checkpoint is restored to restore the applicationto the state prior to the suspensionIn this effort code security inside the cloud is enhancedby exploiting encryption and isolation approaches thatprotects offloaded code from cloud vendors eavesdrop-ping Using probabilistic communication QoS techniquethis is aimed to provide a communication-QoS trade-off For instance the control data (usually small vol-ume) needs highest accuracy while video streaming data(often large volume) requires less communication accu-racy Moreover the authors are optimistic that offeringsecondary tasks such as automatic virus scanning databackup and file sharing in the virtual environment canenhance quality of user experienceAlthough this approach aims to enhance the quality of

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 17: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 17

application execution and augment computation capabil-ity of mobile clients and save energy but responsivenessin interactive applications are likely low due to remote UIexecution Instead of migrating entire application to thecloud it might be more beneficial to utilize some of thelocal mobile resources instead of treating mobile deviceas a dump client Data passing between mobile device andcloud for interactive applications might degrade qualityof experience especially in low-bandwidth intermittentnetworks

bull Virtualized ScreenVirtualized screen [42] is anotherexample of CMA approaches that aims to move the screenrendering process to the cloud and deliver the renderedscreen as an image to the mobile device The authorsaim to enrich the user experience and migrate the screenrendering tasks to the cloud with the assumption thatmajority of computation- and data-intensive processingtake place in the cloud Hence abundant cloud resourcesrsquoexploitation simplifies the CMA system architectureprolongs mobile battery and enhances the interactionand responsiveness of mobile applications toward richuser experience Screen virtualization technique (runningpartial rendering in cloud and rest in mobile dependingon the execution context) is envisioned to optimize userexperience especially for lightweight high-fidelity in-teractive mobile applications that entirely run on localresources Their conceptual proposal aims to enhancevisualization capability of mobile clients mitigate theimpact of hardware and platform heterogeneity and facil-itate porting mobile applications to various devices (egsmartphone laptop and IP TV) with different screensTo reduce the mobile-cloud data transmission a frame-based representation system is exploited to forward thescreen updates from the cloud to the mobile Frame-basedrepresentation system captures and feeds the whole screenimage to the transmission unit This approach updateseach frame based on the previous frame stored insideboth the mobile and cloud However a rich interactiveresponsive GUI needs live streaming of screen imageswhich is impacted by communication latency Althoughthe authors describe optimized screen transmission ap-proaches to reduce the traffic the impact of computationand communication latency is not yet clear as this isa preliminary proposal Moreover utilizing virtualizedscreen method for developing lightweight mobile-cloudapplication is a non-trivial task in the absence of itsprogramming API

bull Cloud-Mobile Hybrid (CMH) ApplicationUnlike appli-cation offloading solutions authors in this proposal [32]introduce a new approach of utilizing cloud resourcesfor mobile users In this effort the authors propose anovel CMH application model in which heavy compo-nents are developed for cloud-side execution whereaslightweight or native codes are developed for mobiledevices execution CMH Applications execution does notneed profiling partitioning and offloading processes andhence produce least computation overhead on mobiledevices Upon successful cloud-side execution the results

are returned back to the mobile for integrating to thenative mobile componentsHowever developing CMH applications is significantlycomplex due to the interoperability and vendor lock-inproblems in clouds and fragmentation issue in mobiles[51] Cloud components designed for a specific cloud arenot able to move to another cloud due to underlying het-erogeneity among clouds Similarly mobile componentsdeveloped for a particular platform cannot be ported todifferent platforms because of heterogeneity Yet isolatingdevelopment of mobile and cloud components createsfurther versioning and integration challengesTo mitigate the complexity of CMH application devel-opments and facilitate portability the authors leverageDomain Specific Language (DSL) [147] [148] A DSLis a programming language with major focus on solvingproblem in specific domains MATLAB23 is a well-knownDSL-based tool for mathematicians A parser takes a DSLscript and converts codes into an in-memory object to beforwarded to various automatic component generatorsThe system needs different code generators for variousmobile and cloud platforms Once the mobile and cloudcomponents are generated the CMH application can beassembled for various mobile-cloud platforms Howeverutilizing DSL-based techniques requires more generaliza-tion efforts to be beneficial in developing all types ofCMH applications

bull microCloud Similar to the CMH frameworkmicroCloud [36] isa modular mobile-cloud application framework that aimsto facilitate mobile-cloud application generation promoteapplication portability minimize the development com-plexity and enhance offline usability in intensive mobile-cloud applications Fulfilling separation of concerns vi-sion skilled programmers independently develop self-contained components which do not have any direct intercommunications with each other Unskilled mobile userscan mash-up (assembling available components to buildcomplex application) these prefabricated components togenerate a complex mobile-cloud application Cloud ven-dors provide infrastructure and platform as cloud servicesto run prefabricated cloud components The main idea inthis proposal is to avoid local execution of the resource-intensive components Hence components are identifiedas cloud mobile and hybrid mobile components areexecutable exclusively on mobile and cloud componentsare strictly developed for cloud server while hybrid com-ponents can either run locally or remotely Hybrid com-ponents have either multiple implementations or a singleimplementation that need a middleware for executionEach component has a triplet of identifier inputoutputparameters and configurationTo alleviate offline usability issue the authors leveragemobile-side queuing and cloud-side caching to main-tain data in case of disconnection Data will be trans-ferred upon reconnection Application is partitioned intocomponents and organized as a directed graph Nodes

23httpwwwmathworkscomproductsmatlab

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 18: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 18

represent components and vertices indicate datacontrolflow Application is divided into three fragments in eachfragment a managing unit called orchestrator executesand maintains componentrsquos mash-up process The outputof each component is forwarded using the pass-by-valuesemantic as an input to the subsequent componentUnlike Elastic Application model the design and im-plementation of components inmicroCloud is statically per-formed in early development phase Thus any improve-ment in resource availability of mobile devices or envi-ronmental enhancement (like bandwidth growth) will notimprove the overall execution ofmicroCloud applicationsSuch inflexibility decreases the application executionperformance and degrades the quality of user experience

bull SmartBoxSmartbox [112] is a self-management on-line cloud-based storage and access management systemdeveloped for mobile devices to expand device storageand facilitate data access and sharing It is a write-once read many times system designed to store personaldata such as text song video and movies which isnot appropriate for large scale computational datasets InSmartbox mobile devices are associated with a shadowstorage to storeretrieve personal data using a uniqueaccount To facilitate data sharing among larger groupof end-users in office or at home a public storage spaceis provisionedSmartbox exploits traditional hierarchical namespacefor smooth navigation and employs an attribute-basedmethod to facilitate data navigation and service queryusing semantic metadata such as the publisher-providermetadata Data navigation and query using tiny keyboardand small screen irk mobile users when inquiring andnavigating stored data in cloud However mobile usersneed always-on connectivity to access online cloud datawhich is not yet achieved and is unlikely to becomereality in near future

bull WhereStoreWhereStore [149] is a location-based datastore for cloud-interacting mobile devices to replicatenecessary cloud-stored mobile data on the phone Themain notion in this effort is that users in different placesdoing various activities need dissimilar types of informa-tion For instance a foreign tourist in Manhattan requiresinformation about nearby places of interest rather than allthe country Hence identifying the location and cachingpredicted data deemed can enhance the system efficiencyand user experience However efficient prediction offuture user location and required data and determiningthe right time for caching data are challenging tasks

bull WukongWukong [150] is a cloud oriented file servicefor multiple mobile devices as a user-friendly and highlyavailable file service The authors provision support ofheterogeneous back-end services such as FTP Mail andGoogle Docs Service in a transparent manner leveraging aservice abstraction layer (SAL) Wukong enables appli-cations to access cloud data without being downloadedinto the local storage of mobile deviceAuthors introduce cache management and pre-fetchmechanisms in different scenarios to increase perfor-

mance while decreasing latency However it cannot al-ways reduce latency due to the bandwidth limitation andIO overhead In operations with long gap between openand read it is beneficial to pre-fetch data from cloudto the device that significantly improves user experienceData security is enhanced via an encryption moduleand bandwidth is saved using a compression moduleWhile compression methods utilized in this proposal isinefficient for multimedia files like image and music itcan compress text and log files noticeablyWe conclude that one of the most effective solutions totackle bandwidth and latency limitations in CMA ap-proaches especially cloud storage is to decrease the vol-ume of data using imminent compression methods Whilevarious compression methods work well on specific filetypes a cognitive or adaptive compression method withfocus on multimedia files can significantly improve thefeasibility of cloud-storage systems

B Proximate Fixed

Researchers have recently proposed CMA approaches inwhich nearby stationary computers are utilized Utilizingnearby desktop computers initiates new generation of servicesto the end-user via mobile device In [26] the authors pro-vide a real scenario in which Ron a patient diagnosed withAlzheimer receives cognitive assistance using an augmented-reality enabled wearable computer The system consists of alightweight wearable computer and a head-up display suchas Google Glass24 equipped with a camera to capture theenvironment and an earphone to send the feedback to thepatient The system captures the scene and sends the image tothe nearby fix computers to interpret the scene in the imageusing the object or face recognition voice synthesizer andcontext-awareness algorithms When Ron looks at a person forfew seconds the personrsquos name and some clue informationis whispered in Ronrsquos ear to help greeting with the personWhen he looks at his thirsty plant or hungry dog the systemreminds Ron to irrigate the plant and feed his dog The nearbyresources are core component of this system to provide low-latency real-time processing to the patient In this part weexplain one of the most prominent proximate fixed efforts asfollows

bull Cloudlet Cloudlet [26] is a proximate immobile cloudconsists of one or several resource-rich multi-core Gi-gabit Ethernet connected computer aiming to augmentneighboring mobile devices while minimizing securityrisks offloading distance (one-hop migration from mo-bile to Cloudlet) and communication latency Mobiledevice plays the role of a thin client while the intensivecomputation is entirely migrated via Wi-Fi to the nearbyCloudlet Although Cloudlet utilizes proximate resourcesthe distant fixed cloud infrastructures are also accessiblein case of Cloudlet scarcity The authors employ a decen-tralized self-managed widely-spread infrastructure builton hardware VM technology Cloudlet is a VM-based

24httpwwwgooglecomglassstart

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 19: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 19

Fig 7 Cloudlet-based Resource-Rich Mobile Computing Life Cycle

offloading system that can significantly shrink the impactof hardware and OS heterogeneity between mobile andCloudlet infrastructuresTo reduce the Cloudlet management and maintenancecosts while increasing security and privacy of bothCloudlet host and mobile guest a method called ldquotran-sient Cloudlet customizationrdquo is deployed which useshardware VM technology It enables Cloudlet customiza-tion prior to the offloading and performs Cloudlet restora-tion as a post-offloading cleanup process to restore thehost to its original software stake The VM encapsulatesthe entire offloaded mobile environment (data state andcode) and separates it from the host permanent softwareHence feasibility of deploying Cloudlet in public placessuch as coffee shops airport lounge and shopping mallsincreasesUnlike CloneCloud and Virtual Execution Environmentefforts that migrate the entire mobile OS clone to thecloud Cloudlet assumes that the entire OS clone existsand is preloaded in the host and runs on an isolatedVM In mobile side instead of creating the VM ofthe entire mobile application and its memory stack thesystems encapsulates a lightweight software interface ofthe intensive components called VM overlayThe VM overall offloading performance is further en-hanced by exploiting Dynamic VM Synthesis (DVMS)method since its performance solely depends onthe mobile-Cloudlet bandwidth and cloudlet resourcesDVMS assumes that the base VM is already availablein the target Cloudlet and user can find the match-

ing execution environment (VM base) among silo ofnearby Cloudlets Upon discovery and negotiation of theCloudlet the DVMS offloads the VM overlay to theinfrastructure to execute launch VM (base + overlay)Henceforth the offloaded code starts execution in thestate it was paused Upon completion of Cloudlet execu-tion the VM residue is created and sent back to the mobiledevice In the Cloudlet the VM is discarded as a post-offloading cleanup process to restore the original Cloudletstate In mobile side the results will be integrated tothe application and local execution will be resumed Topresent a clear understanding of the overall process thesequence diagram of Cloudlet-based resource-rich mobilecomputing is depicted in Figure 7Despite the noticeable offloading improvements in theCloudlet its success highly depends on the existence ofplethora of powerful Cloudlets containing popular mobileplatformsrsquo base VM Encouraging individual owners todeploy such Cloudlets in the absence of monetary incen-tives is an issue that must be addressed before deploymentin real scenarios Although energy efficiency securityand privacy and maintenance of Cloudlet are widelyacceptable further efforts are required to protect theoverall CMA process Moreover few minutes offloadinglatency in Cloudlet is unacceptable to users [151]

C Proximate Mobile

Recently several researchers [24] [45] [122] [152]ndash[155]propose CMA approaches in which nearby mobile deviceslend available resources to other mobile clients for execution

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 20: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 20

Fig 8 MOMCC Concept

of resource-intensive tasks in distributed manner Utilizingsuch resources can enhance user experience in several realscenarios such as Optical Character Recognition (OCR) andnatural language processing applications The feasibility ofutilizing nearby mobile devices is studied in [24] where Petera foreign tourist visiting a Korean exhibition and finds interestin an exhibit but cannot understand the Korean descriptionHe can take a photo of the manuscript and translate it usingthe OCR application but his device lacks enough computationresources Although he can exploit the Internet web servicesto translate the text the roaming cost is not affordable to himHence he leverages a CMA solution by utilizing computationresources of nearby mobile devices to complete the task Someof the CMA efforts whose remote resources are proximatemobile devices are explained as follows

bull MOMCC Market-Oriented Mobile Cloud Computing(MOMCC) [45] is a mobile-cloud application frame-work based on Service Oriented Architecture (SOA)that harnesses a cluster of nearby mobile devices torun resource-intensive tasks In MOMCC mobile-cloudapplications are developed using prefabricated buildingblocks called services developed by expert programmersService developers can independently develop variouscomputation services and uploaded them to a publiclyaccessible UDDI (Universal Description Discovery andIntegration) such as mobile network operatorsServices are mostly executed on large number of smart-phones in vicinity which can share their computationresources and earn some money To enhance resourceavailability and elasticity distant stationary cloud re-sources are also available if nearby resources are in-sufficient In order to become an IaaS (Infrastructureas a Service) provider mobile devices register with theUDDI and negotiate to host certain services after secureauthentication and authorization Mobile users at runtimecontact UDDI to find appropriate secure host in vicinityto execute desired service on payment The collected rev-

enue is shared between service programmer applicationdeveloper UDDI and service host for promotion andencouragement Figure 8 depicts the MOMCC conceptHowever MOMCC is a preliminary study and its overallperformance is not yet evident Several issues are requiredto be addressed prior to its successful deployment inreal scenarios Executing services on mobile devices is achallenging task considering resource limitation securityand mobility Also an efficient business plan that cansatisfy all engaging parties in MOMCC is lacking anddemand future efforts MOMCC can provide an incomesource for mobile owners who spend couple of hundreddollars to buy a high-end device In addition faultyresource-rich mobile devices that are able to functionaccurately can be utilized in MOMCC instead of beinge-waste

bull Hyrax Hyrax [152] is another CMA approach that ex-ploits the resources from a cluster of immobile smart-phones in vicinity to perform intense computationsHyrax alleviates the frequent disconnections of mobileservers using fault tolerance mechanism of Hadoop Sim-ilar to MOMCC and Cloudlet due to resource limita-tions of smartphone servers the accessibility to distantstationary clouds is also provisioned in case the nearbysmartphone resources are not sufficient However Hyraxdoes not consider mobility of mobile clients and mobileservers Hence deployment of Hyrax in real scenariosmay become less realistic due to immobilization of mo-bile nodes Lack of incentive for mobile servers alsohinders Hyrax successHyrax is a MCC platform developed based on Hadoop[156] for Android smartphones In developing Hyraxthe MapReduce [157] principles are applied utilizingHadoop as an open source implementation of MapRe-duce MapReduce is a scalable fault-tolerant program-ming model developed to process huge dataset over acluster of resources Centralized server in Hyrax runs two

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 21: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 21

client side processes of MapReduce namely NameNodeand JobTracker processes to coordinate the overall com-putation process on a cluster of smartphones In smart-phone side two Hadoop processes namely DataNodeand TaskTracker are implemented as Android servicesto receive computation tasks from the JobTracker Smart-phones are able to communicate with the server and othersmartphones via 80211g technologyNevertheless the cloud storage connectivity in Hyrax ismissing It demands several gigabytes of local storage tostore data and computation Hence user cannot accessdistributed data over the Internet or Ethernet The authorutilizes the constant historical multimedia data to avoidfile sharing Hence it is less beneficial for interactiveand event-oriented applications whose data frequentlychanges over the execution and also data-intensive appli-cations that require huge database The overall overheadin Hyrax is high due to the intensity of Hadoop algorithmwhich runs locally on smartphones

bull Virtual Mobile Cloud Computing (VMCC)Researchersin [24] aim to augment computing capabilities of stablemobile devices by leveraging an ad-hoc cluster of nearbysmartphones to perform intensive computing with min-imum latency and network traffic while decreasing theimpact of hardware and platform heterogeneity Duringthe first execution required components (proxy creationand RPC support) are added to the application code tobe used for offloading the modified code will remain forfuture offloading For every application the system de-termines the number of required mobile servers securityand privacy requirements and offloading overhead Thesystem continuously traces the number of total mobileservers and their geolocation to establish a peer-to-peercommunication among them Upon decision making theapplication is partitioned into small codes and transferredto the nearby mobile nodes for execution The results willbe reintegrated back upon completionHowever several issues encumber VMCCrsquos successFirstly this solution similar to Hyrax is not suitablefor a moving smartphone since the authors explicitlydisregard mobility trait of mobile clients Secondly everycomputing job is sent to exactly one mobile node so theoffloading time and overhead will be increased when theserving node leaves the cluster Thirdly the offloadinginitiation might take long since the offloadingrsquos overallperformance highly depends on the number of availablenearby nodes insufficient number of mobile nodes defersoffloading Finally in the absence of monetary incentivefor mobile nodes the likelihood of resource sharingamong resource-constraint mobile devices is low

D Hybrid

Hybrid CMA efforts are budding [46] [143] [158] to opti-mize the overall augmentation performance and researchersareendeavoring to seamlessly integrate various types of resourcesto deliver a smooth computing experience to mobile end-users For instance mCloud [159] is an imminent proposal to

integrate proximate immobile and distant stationary computingresources Authors are aiming to enable mobile-users to per-form resource-intensive computation using hybrid resources(integrated cloudlet-cloud infrastructures) Hybrid solutionsaim to provide higher QoS and richer interaction experienceto the mobile end-users of real scenarios explained in previousparts For instance in the foreign tourist example the imagecan be sent to the nearby mobile device of a non-native localresident for processing When the processing fails due to lackof enough resources the picture can be forwarded to the cloudwithout Peter pays high cost of international roaming (Petermay pay local charge)

We review some of the available hybrid CMAs as follows

bull SAMI SAMI (Service-based Arbitrated Multi-tier In-frastructure for mobile cloud computing) [46] proposesa multi-tier IaaS to execute resource-intensive compu-tations and store heavy data on behalf of resource-constraint smartphones The hybrid cloud-based infras-tructures of SAMI are combination of distant immobileclouds nearby Mobile Network Operators (MNOs) andcluster of very close MNOs authorized dealers depictedin Figure 9 The compound three level infrastructures aimto increase the outsourcing flexibility augmentation per-formance and energy efficiency The MNOrsquos revenue ishiked in this proposal and energy dissipation is preventedNearby dealers can be reached by Wi-Fi MNOrsquos can beaccessed either directly via cellular connection or throughdealers via Wi-Fi and broadband Connection is estab-lished via cellular network to contact distant stationaryclouds The cluster of nearby stationary machines (MNOdealers located in vicinity) performs latency-sensitiveservices and omits the impact of network heterogeneitySAMI leverages Wi-Fi technology to conserve mobileenergy because it consumes less energy compared tothe cellular networks [116] In case of nearby resourcescarcity or end-user mobility the service can be executedinside the MNO via cellular network However if theresources in MNO are insufficient the computation canbe performed inside the distant immobile cloudThe resource allocation to the services is undertaken byarbitrator entity based on several metrics particularlyresource requirements latency and security requirementsof varied services The arbitrator frequently checks andupdates the service allocation decision to ensure highperformance and avoids mismatchTo enhance security of infrastructures SAMI employscomparatively reliable and trustworthy entities namelyclouds MNOs and MNO trusted dealers MNOs havealready established reputation-trust among mobile usersand can enforce a strict security provisions to establishindirect trust between dealers and end users ensuring thatuserrsquos security and privacy will not be violated SAMI ap-plication development framework facilitates deploymentof service-based platform-neutral mobile applications andeases data interoperability in MCC due to utilization ofSOAHowever SAMI is a conceptual framework and deploy-

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 22: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 22

Fig 9 SAMI A Multi-Tier Cloud-based Infrastructure Model

ment results are expected to advocate its feasibility SAMIimposes a processing overhead on MNOs due to con-tinuous arbitration process Deployment managementand maintenance costs of SAMI are also high due tothe existence of various infrastructure layers Moreoverthough the authors discuss the monetary aspects of theproposal a detailed discussion of the business plan ismissing for example in what scenario resource outsourc-ing is affordable for the mobile application How doesincome should be divided among different entities to besatisfactory

bull MOCHA In MOCHA [158] authors propose a mobile-cloudlet-cloud architecture for face recognition applica-tion using mobile camera and hybrid infrastructures ofnearby Cloudlet and distant immobile cloud Cloudletis a specific cheap cluster of computing entities likeGPU (Graphics Processing Unit) capable of massivelyprocessing data and transactions in parallel Cloudletsare able to be accessed via heterogeneous communicationtechnologies such as Wi-Fi Bluetooth and cellular Themobile often access processing resources via Cloudletrather than directly connecting to the cloud unless ac-cessing cloud resources bears lower latencyCloudlet receives the smartphones intensive computationtasks and partitions them for distribution between it-self and distant immobile clouds to enhance QoS [26]MOCHA leverages two partitioning algorithms fixedand greedy In the fixed algorithm the task is equallypartitioned and distributed among all available computingdevices (including Cloudlet and cloud servers) whereas

in greedy algorithm the task is partitioned and distributedamong computing devices based on their response timesthe first partition is sent to the quickest device while thelast partition is sent to the slowest device The responsetime of the task partitioned using greedy approach issignificantly better than fixed especially when Cloudletserver is utilized in augmentation process and largenumber of clouds with heterogeneous response time existHowever smartphones in MOCHA require prior knowl-edge of the communication and computation latency ofall available computing entities (Cloudlet and all availabledistant fixed clouds) which is a resource-hungry and time-consuming task

VI CMA PROSPECTIVES

People dependency to mobile devices is rapidly increasing[89] [160] and smartphones have been using in several crucialareas particularly healthcare (tele-surgery) emergency anddisaster recovery (remote monitoring and sensing) and crowdmanagement to benefit mankind [161]ndash[163] However intrin-sic mobile resources and current augmentation approaches arenot matching with the current computing needs of mobile-users and hence inhibit smartphonersquos adoption Upon slowprogress of hardware augmentation the highly feasible solu-tion to fulfill people computing needs is to leverage CMAconcept This Section aims to present set of guidelines forefficiency adaptability and performance of forthcoming CMAsolutions We identify and explain the vital decision makingfactors that significantly enhance quality and adaptability offuture CMA solutions and describe five major performancelimitation factors We illustrate an exemplary decision makingflowchart of next generation CMA approaches

A CMA Decision Making Factors

These factors can be used to decide whether to performCMA or not and are needed at design and implementationphases of next generation CMA approaches We categorizethe factors into five main groups of mobile devices contentsaugmentation environment user preferences and requirementsand cloud servers which are depicted in Figure 11 andexplained as follows

1) Mobile Devices From the client perspective amountof native resources including CPU memory and storage isthe most important factor to perform augmentation Alsoenergy is considered a critical resource in the absence of long-spanning batteries The trade-off between energy consumedbyaugmentation and energy squandered by communication is avital proportion in CMA approaches [73] Device mobility andcommunication ability (supporting varied technologies suchas 2G3GWi-Fi) are other metrics that are important in theoffloading performance

2) ContentsAnother influential factor for CMA decisionmaking is the contentsrsquo nature The code granularity andsize as well as data type and volume are example attributesof contents that highly impact on the overall augmentationprocess Hence the augmentation should be performed con-sidering the nature and complexity of application and data

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 23: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 23

Fig 10 Comparison of CMA Approaches

For instance latency sensitive small data are efficient to beprocessed locally whereas sensitive big data are encouraged tobe stored in a large reliable cloud storage Similarly offloadinga coarse-grained large code to a distant fixed cloud via a lowbandwidth network is not feasible

3) Augmentation EnvironmentMobile computing is a het-erogeneous environment comprised of non-uniform mobilenodes communication technologies and resources One of themost influential environment-dependent factors is the wirelesscommunication medium in which majority of communicationstake place Wireless is an intermittent unreliable risky andblipping medium with significant impact on the quality ofaugmentation solutions The overall performance of a low costhighly available and scalable CMA approach is magnificentlyshrunk by the low quality of communication medium andtechnologies Selecting the most suitable technology consider-ing the factors like required bandwidth congestion utilizationcosts and latency [164] is a challenge that affects qualityof augmentation approaches in wireless domains Wirelessmedium characteristics impose restrictions when specifyingremote servers at design time and runtime

Moreover dynamism and rapidly changing attributes of theruntime environment noticeably impact on augmentation pro-cess and increase decision making complexity Augmentationapproaches should be agile in dynamic mobile environmentand instantaneously reflect to any change For example usermovement from high bandwidth to a low bandwidth network

receding from the network access point and rapidly changingavailable computing resources complicate CMA process

4) User Preferences and RequirementsEnd-usersrsquo physi-cal and mental situations individual and corporate preferencesand ultimate computing goals are important factors that affectoffloading performance Some users are not interested toutilize the risky channel of Internet while others may demandaccessing cloud services through the Internet Hence usersshould be able to modify technical and non-technical spec-ifications of the CMA system and customize it according totheir needs For example user should be able to alter degreeofacceptable latency against energy efficiency of an applicationexecution Selecting the most appropriate resource amongavailable options can also enhance overall user experience

5) Cloud Servers As explained CMA approaches canleverage various types of cloud resources to enhance com-puting capabilities of mobile devices Therefore the overallperformance and credibility of the augmentation approacheshighly depend on the cloud-based resourcesrsquo characteristicsPerformance availability elasticity vulnerability tosecurityattacks reliability (delivering accurate services basedonagreed terms and conditions) cost and distance are majorcharacteristics of the cloud service providers used for aug-menting mobile devices

Utilizing clouds to augment mobile devices notably reducesthe device ownership cost by borrowing computing resourcesbased on pay-as-you-use principle Such elastic cost-effective

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 24: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 24

Fig 11 Critical Factors in CMA Decision Making

reliable and relatively trustworthy resources are embraced bythe scholars industrial organizations and end-users towardsflourishing CMA approaches

B Performance Limitation Factors

Performance of varied CMA solutions is impacted by sev-eral factors We describe fix major performance limitationfactors as follows

1) HeterogeneityMCC is a highly heterogeneous envi-ronment comprised of three diversified domains of mobilecomputing cloud computing and networking Although het-erogeneity can provide flexibility to the mobile users byproviding selection alternatives it breeds several limitationsand challenges especially for developing multi-tier CMA-based applications [51] Dissimilar mobile platforms suchasAndroid iOS Symbian and RIM beside diverse hardwarecharacteristics of mobile device inhibit data and applicationportability among varied mobile devices Portability is the abil-ity to migrate code and data from one device to another withnoless modification and change [165] Existing heterogene-ity in cloud computing including hardware platform cloudservice policy and service heterogeneity originates challengessuch as portability and interoperability and fragment the MCCdomain

Network heterogeneity in MCC is the composition of var-ious wireless technologies such as Wi-Fi 3G and WiMAXMobility among varied network environments intensifies com-munication deficiencies and stems complex issues like signalhandover [125] Inappropriate decision making during thehandover process like (i) less appropriate selection of networktechnology among available candidates and (ii) transferring thecommunication link at the wrong time increases WAN latency

and jitter that degrade quality of mobile cloud services Con-sequently delay-sensitive content and services are degraded[166] and adoption of CMA approaches are hindered

2) Data Volume Ever-increasing volume of digital con-tents [85] significantly impacts on the performance of CMAapproaches in MCC Current wireless infrastructures and tech-nologies fail to efficiently fulfill the networking requirementsof CMA approaches Storing such a huge data in a single ware-house is often impossible and demands data partitioning anddistributed storage that not only mitigates data integrityandconsistency but also makes data management a pivotal needin MCC [167] Applying a single access control mechanismfor relevant data in various storage environments is anotherchallenging task that impacts on the performance and adoptionof CMA solutions in MCC

3) Round-Trip LatencyCommunication and computationlatency is one of the most important performance metrics ofmobile augmentation approaches especially when exploitingdistant cloud resources In cellular communications distancefrom the base station (near or far) and variations in bandwidthand speed of various wireless technologies affect the perfor-mance of augmentation process for mobile devices Moreoverleveraging wireless Internet networks to offload content tothedistant cloud resources creates a bottleneck Latency adverselyimpacts on the energy efficiency [73] and interactive response[168] of CMA-based mobile applications due to excessiveconsumption of mobile resources and raising transmissiondelays

Recently researches [169] [170] are emerging toward de-creasing the networking overhead and facilitating mobility(both node and code mobility) in cloud-based offloadingapproaches For example Follow-Me Cloud [169] aims atenabling mobility of network end-points across different IP

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 25: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 25

subnets The authors employ the concept of identifier andlocator separation of edge networks using OpenFlow-enabledswitches Leveraging the Follow-Me Cloud mobile nodescan move among access networks without being notified ofany change or session disruption All corresponding nodesthat have been communicating with the mobile node cancontinue their communication without interruption When thenode migrates its old IP turn to identifier and its new IPaddress becomes locator address so that all other nodes cankeep communication with the moving node However foreach packet traveling tofrom the mobile node there is anoverhead of manipulating the locatoridentifier values Futureimprovement and optimization efforts will enhance the CMAsystemsrsquo performance

In cloud side computation latency significantly impacts onthe application responsiveness Researchers study the impactof cloud computation performance on the execution timeand vindicate 12X reduction in performance time violation[171] Thus the increased latency degrades the quality ofuser experience and adversely impacts on the user-perceivedperformance of CMA solutions

4) Context Management and ProcessingPerformance ofCMA approaches is noticeably degraded by lack of sufficientaccurate knowledge about the runtime environment Contem-porary mobile devices are capable of gathering extensive con-text and social information such as available remote resourcesnetwork bandwidth weather conditions and usersrsquo voice andgestures from their surrounding environment [137] [138]Butstoring managing and processing large volume of contextinformation (considering MCC environmentrsquos dynamism andmobile devicesrsquo mobility) on resource-constraint smartphonesare non-trivial tasks

5) Service Execution and DeliverySLA as a formalcontract between service consumer and provider enforcesresource-level QoS (eg memory capacity compute unit andstorage) against a fee which is not sufficient for mobile usersin highly dynamic wireless MCC environment User-perceivedperformance in MCC is highly affected by the quality of cloudcomputations wireless communications and local executionHence varied service providers including cloud vendorswireless network providers and mobile hardware and OSvendors need to collaborate and ensure acceptable level ofQoS For successful CMA approaches comprehensive real-time monitoring process is expected to ensure that engagingservice vendors are delivering required services in acceptablelevel based on the accepted SLA

C CMA Feasibility

Although CMA is beneficial and can saves resources [40]several questions need to be addressed before CMA can beimplemented in real scenarios For instance is CMA alwaysfeasible and beneficial Can CMA save local resources andenhance user experience What kind of cloud-based resourcesshould be opted to achieve the superior performance

Vision of future CMA proposals will be realized by accuratesensing and acquiring precise knowledge of decision makingfactors like user preferences and requirements augmentation

environment and mobile devices which are explained inprevious part A decision making system similar to those usedin [25] [33] [49] analyzes these vital factors to determinethe augmentation feasibility and specifies if augmentationcanfulfill mobile computation requirements and enrich qualityof user experience Figure 12 illustrates a possible decisionmaking flow of future CMA approaches

Availability of mobile resources to manage augmentationprocess and volume of cloud resources to provision requiredresources significantly impact on the quality of augmentation[9] Similarly user preferences limitations and requirementsaffect the augmentation decision making For instance if aug-mentation is not permitted by users the application executionand data storage should be performed locally without beingoffloaded to a remote server(s) or be terminated in the absenceof enough local resources Similarly augmentation processcan be terminated if the execution latency of delay-sensitivecontent is sharply increased quality of execution is noticeablydecreased or security and privacy of users is violated [19]

Furthermore usefulness of CMA approaches highly de-pends on the execution environment Offloading computationand mobile-cloud communication ratio distance from mobileto the cloud network technologies and coverage availablebandwidth traffic congestion deployment cost and even na-ture of augmentation tasks alter usefulness of the CMA ap-proaches [40] For instance performing an offloading methodon a data-intensive application (eg applying a graphical filteron large number of high quality images) in a low-bandwidthnetwork imposes large latency and significantly degrades userexperience which should be avoided Similarly migrating aresource-hungry code to an expensive remote resource canbe unaffordable practice Suppose in a sample augmentationapproachRC is the total native resources consumed duringaugmentationRM is the total native resources consumedfor maintenance andRS is the total resources conserved inaugmentation process Explicitly for a feasible augmentationapproachRC + RM ltlt RS However in some traditionalaugmentation approaches the left side of the equation exceedsthe right which is not effective in augmenting resource-poormobile devices [19]

VII O PEN CHALLENGES

In this Section we highlight some of the most importantchallenges in deploying and utilizing CMA approaches as thefuture research directions

A Reference Architecture for CMA Development

Recently researchers leverage dissimilar structures andtechniques in utilizing cloud resources to augment compu-tation capabilities of mobile devices Also different effortsfocus on varied types of mobile applications particularlymultimedia intense games image processing and workflowprocessing applications

Such diffusion scatters CMA development approaches andincrease adaptability challenges of CMA solutions In theabsence of a reference architecture and unified CMA solutionvarious CMA approaches need to be integrated to all mobile

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 26: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 26

Fig 12 An Exemplary Decision Making Flow of Future CMA Approaches

OSs to serve multi-dimensional needs of various mobile userswhich is a not-trivial task The reference architecture is ex-pected to be generic enough to be deployed in family of CMAapproaches

B Autonomic CMA

CMA approaches are drastically increasing volume of dis-tributed mobile content in a horizontally heterogeneous mobilecloud computing environment [120] Exploiting heterogeneouscommunication technologies to employ diverse cloud-basedinfrastructures for augmenting a plethora of dissimilar mobiledevices is significantly intensifying complexity and manage-ment Employing lateral solutions and controlling the mobilephones using outside entities (eg third party managementsystems) to deal with such complexity might further amplifycomplexity A feasible alternative to mitigate such complexityis to develop autonomic self-managing -healing -optimizingand -protecting CMA approaches [172] able to adapt toenvironment dynamism and hide complexity

C Application Mobility Provisioning

Enabling continuous and consistent mobility in CMA mod-els (especially proximate mobile and hybrid) to provisionubiquitous convenient on-demand network access to cloud-based computing resources is a vital challenge Seamless codemobility in CMA models is more challenging compared tothe traditional augmentation approaches because in CMA

approaches service providers and consumers can move duringthe augmentation process which intensifies the code mobility[173] Therefore communication disruption and intermittencycan cause several challenges especially dismissal of always-on connectivity excessive consumption of limited mobileresources and frequent interruption of application executionwhich decreases quality of computing services and degradesquality of mobile user experience [72] Also it levies re-dundant costs on cloud-mobile users and inhibits reliabilityof CMA models Hence alleviating such difficulties usingWeb advancements [170] and imminent lightweight cognitivemobility management systems with least signal traffic andlatency can significantly enhance the ubiquitous connectivityand increase the positive impact of CMA

D Computing and Temporal Cost of Mobile Distributed Ex-ecution

Noticeable computation and communication cost of migrat-ing tasks from the mobile device to the remote servers andreceiving the results is another challenges of CMA approachesin MCC which is intensified by mobility and wireless commu-nication constraints Although researchers [45] [152] endeavorto reduce the distance of mobile devices and service providersby leveraging nearby mobilefixed computing devices severalmechanisms particularly resource discovery and allocationservice consumer and provider mobility management anddistributed runtime are required to realize the CMAs visionAccurate allocation of resources to the mobile computationtasks demands comprehensive knowledge about structure andperformance features of available service providers and re-source requirements of mobile computation tasks Thus QoS-aware scheduling efforts such as [143] are necessary to en-hance the CMA usability

E Seamless Communication

Maintaining a continuous communication between mobileservice consumers and mobilefixed service providers in in-termittent heterogeneous wireless environment is a non-trivialtask User mobility and wireless disconnection highly im-pact on resource utilization ratio When the mobile serviceproviders and consumers loss communication link due tomobility-made prolonged distance the service consumer re-quires to either performing local execution or re-initiating aug-mentation process Similarly the resource-constraint mobileserver consumes its scarce resources for processing an orphancomputation whose results are ineffectual after losing thecommunication link Potential solutions may transfer partially-completed tasks to a nearby node or initiate parallel executionon third device before disconnection or cache results for futurereferences

F Multipoint Data Bridging

Unlike traditional offloading methods that require point-to-point code and data migration and processing CMA ap-proaches require multipoint data migration and interoperation

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 27: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 27

to achieve the maximum benefits from the distributed het-erogeneous infrastructures in mobile-cloud ecosystem Con-necting heterogeneous systems (based on wired or wireless)understanding geographical information resources and ex-changing data betweenacross two or more heterogeneoussystems [174] are the main issues in CMA system whichdemand arousing data interoperation techniques in multi-domains mobile cloud environment The inward heterogeneousarchitectures and data structures of mobile devices and cloudsystems with different APIs can exemplify the intensity ofmultipoint data bridging challenge [23] Hence offloadingcomputational tasks from a mobile device to a cloud perform-ing computational interoperation among varied clouds (forcostand performance concerns) and pushing results to the mobiledevice become challenging tasks [175] Therefore multipointdata bridging in dynamic heterogeneous environment remainsas a future research direction to realize accessing interpretingprocessing sharing and synchronizing distributed contents

G Distributed Content Management

Rapid growth in digital contents and increasing depen-dency of mobile users to cloud infrastructures impede contentmanagement for mobile users Researchers distribute codeand data among heterogeneous nearby and distant resourcesvia different communication technologies to optimize CMAprocess Although executing complex heavy applications andaccessing large data volume are facilitated managing hugevolume of distributed contents for smartphone users is notstraightforward Therefore enabling mobile users to efficientlylocate access update and synchronize highly distributedcontents requires future research and developments

H SeamlessLightweight CMA

Developing lightweight mobile computing augmentationapproaches to increase quality of mobile user experienceand to develop CMA system independent of any particularsituation is a significant challenge in mobile cloud environ-ment Offloading bulk data in limited wireless bandwidthand VM initiation migration and management in a secureand confidential manner are particular tasks in CMA systemthat noticeably increase overall execution time intensify theaugmentation latency and decrease the quality of mobileuser experience CMA approaches are generally hosted andexecuted inside the mobile devices to conserve their localresources and hence need to avoid excessive resource hungrytransactions

A feasible approach to decrease the volume of digitalcontents mdashin limited bandwidth networks mdashis to utilizeeffective efficient data compression methods Available com-pression techniques are unlikely efficient considering structureof current multimedia files Moreover approaches like Par-avirtualiztion [176] as a lightweight virtualization techniquecan reduce the overhead by partially emulating the OS andhardware Paravirtualization approaches virtualize onlypartsof the hardware required for computing Thus the mobile-side VM creation overhead is diminished and the impact ofVM migration on network is reduced Therefore realizing

lightweight CMA approaches demands lightweight computa-tion and communication techniques (particularly in virtualiza-tion data compression and encryption methods) to reduceintra-system correspondence data volume and IO tasks

I Security in Mobile Cloud

One of the most challenging aspects of CMA is protectingoffloaded code and data inside the cloud While securing con-tents inside the mobile consumes huge resources offloadingplain contents through insecure wireless medium and storingplain data inside the cloud highly violates user security andprivacy Despite of the large number of research and develop-ment in establishing trust in cloud [177]ndash[180] security andprivacy is still one of the major user concerns in utilizingcloud resources that impede CMA deployment Addressingsuch crucial needs by employing a novel lightweight secu-rity algorithm in mobile side and a set of robust securitymechanisms in cloud demand future efforts to promote CMAamong smartphone users In privacy aspects though recentsocial behaviors of users in social websites such as Facebookand tweeter advocates that large community of users partiallyforfeit privacy they still need certain degree of robust privacyto protect their confidential clandestine data

J Live Virtual Machine Migration

Live migration of VMs between distributed cloud-basedservers (especially for distant servers) is a crucial requirementin successfully adopting CMA solutions in MCC consideringwireless network bandwidth and intermittency and mobilitylimitations When a mobile user moves to a place far fromthe offloaded contents (code or data) the enlarged distanceincreases access latency and degrades user-observed applica-tion performance Hence mobilizing the running VM alongwith the mobile service consumer without perceivable serviceinterruption becomes essential to avoid user experience degra-dation However sharp growing computation and data volumein blipping wireless environment intensifies live migration ofVM Therefore efforts similar to VMware vMotion [181] and[121] [182] are necessary to optimize VM migration in MCCReducing computation complexity and overhead energy datavolume and communication cost are critical in low-latencylow-cost migration of VM in MCC

Furthermore after successful live VM migration it is essen-tial to ensure that migrated VM is seamlessly accessible viainitial IP address when the VM changes its physical machineFuture efforts similar to [183] [184] and LISP [185] areneeded to realize the vision of seamless access to migratingVM in MCC

VIII C ONCLUSIONS

Augmenting computing capabilities of mobile devices es-pecially smartphones using cloud infrastructures and principlesis an emerging research area The ultimate goal of CMAsolutions is to realize the vision of unrestricted functionalitystorage and mobility regardless of underlying devices andtechnologiesrsquo constraints Elasticity availability and security

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 28: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 28

are considered as the most significant characteristics of remotecomputing resources Augmenting smartphones significantlyimproves their usability and adoption in various critical areassuch as healthcare emergency handling disaster recoveryand crowd management Enterprise organizations also leveragesuch value-added services to enhance quality of servicesdelivering to their end-users However several communicationand networking challenges particularly live VM migrationseamless back-end and front-end mobility handover contextawareness and location management decelerate adoption ofcloud-based augmentation solutions in MCC Similarly irrec-oncilable and unpredictable bandwidth jitter communicationdelay inconsistent network throughput and non-elastic wire-less spectrum and infrastructures encumber upfront planningfor CMA

In this paper we presented a comprehensive survey onmobile augmentation process and reviewed several CMA ap-proaches We described deficiencies of current mobile devicesas a motivation for augmentation Comparing surrogates as tra-ditional servers with contemporary cloud-based infrastructuresadvocates adaptability of cloud computing and principles asan appropriate back-end technology in mobile augmentationdomain Cloud resources in this domain are manifested asdistant immobile clouds proximate mobile and immobile com-puting entities and hybrid combination of resources CMAapproaches not only enhance processing energy and storagecapabilities of mobile devices but also amend data safetyand security data ubiquity accessibility and user interfacewhile facilitate distributed application programming Despiteof significant CMA approaches advantages several disadvan-tages hinder their adoption in mobile computing Several openchallenges particularly autonomic CMA seamless applicationmobility distributed data management and multipoint databridging are the most prominent open challenges that demandfuture efforts

REFERENCES

[1] M Weiser ldquoThe computer for the 21st CenturypdfrdquoIEEE PervasiveComputing vol 1 no 1 pp 19ndash25 2002

[2] M Satyanarayanan ldquoMobile computing wherersquos the tofurdquo ACMSIGMOBILE Mobile Computing and Communications Review vol 1no 1 pp 17ndash21 1997

[3] S Abolfazli Z Sanaei and A Gani ldquoMobile Cloud Computing AReview on Smartphone Augmentation Approachesrdquo inProc WSEASCISCOrsquo12 Singapore 2012

[4] M Othman and S Hailes ldquoPower conservation strategy for mobilecomputers using load sharingrdquoACM SIGMOBILE Mobile Computingand Communications Review vol 2 no 1 pp 44ndash51 Jan 1998

[5] A Rudenko P Reiher G J Popek and G H Kuenning ldquoSavingportable computer battery power through remote process executionrdquoACM SIGMOBILE Mobile Computing and Communications Reviewvol 2 no 1 pp 19ndash26 1998

[6] M Satyanarayanan ldquoPervasive computing vision and challengesrdquoIEEE Personal Communications vol 8 no 4 pp 10ndash17 2001

[7] J Flinn D Narayanan and M Satyanarayanan ldquoSelf-tuned remote ex-ecution for pervasive computingrdquo inProc HotOSrsquo01 Krun Germany2001 pp 61ndash66

[8] J Flinn P SoYoung and M Satyanarayanan ldquoBalancingperformanceenergy and quality in pervasive computingrdquo inProc IEEE ICDCS rsquo02Vienna Austria 2002 pp 217ndash226

[9] R K Balan ldquoSimplifying cyber foragingrdquo Doctorate Thesis CarnegieMellon University 2006

[10] H-I Balan R and Flinn J and Satyanarayanan M andSSinnamohideen and Yang ldquoThe Case for Cyber Foragingrdquo inProc ACM SIGOPS EW10 Saint Malo France 2002 pp 87ndash92

[11] R K Balan D Gergle M Satyanarayanan and J Herbsleb ldquoSimpli-fying cyber foraging for mobile devicesrdquo inProc ACM MobiSysrsquo07Puerto Rico 2007 pp 272ndash285

[12] R K Balan M Satyanarayanan S Park and T Okoshi ldquoTactics-basedremote execution for mobile computingrdquo inProc ACM MobiSysrsquo03San Francisco California USA 2003 pp 273ndash286

[13] S Goyal and J Carter ldquoA lightweight secure cyber foraging infrastruc-ture for resource-constrained devicesrdquo inProc MCSArsquo04 Lake DistrictNational Park UK 2004 pp 186ndash195

[14] M D Kristensen ldquoEnabling cyber foraging for mobile devicesrdquo inProc MiNEMArsquo7 Magdeburg Germany 2007 pp 32ndash36

[15] M D Kristensen and N O Bouvin ldquoDeveloping cyber foragingapplications for portable devicesrdquo inProc 7th IEEE Intrsquol ConfPolymers and Adhesives in Microelectronics and Photonicsand 2ndIEEE Intrsquo Interdisciplinary Conf PORTABLE-POLYTRONIC 2008pp 1ndash6

[16] Y-Y Su and J Flinn ldquoSlingshot deploying stateful services in wirelesshotspotsrdquo inProc ACM MobiSysrsquo05 Seattle Washington USA 2005pp 79ndash92

[17] X Gu and K Nahrstedt ldquoAdaptive offloading inference for deliveringapplications in pervasive computing environmentsrdquo inProc IEEEPerCom rsquo03 Dallas-Fort Worth Texas USA 2003

[18] M D Kristensen ldquoScavenger Transparent development of efficientcyber foraging applicationsrdquo inProc Percomrsquo10 Mannheim Germany2010 pp 217ndash226

[19] M Sharifi S Kafaie and O Kashefi ldquoA Survey and Taxonomy ofCyber Foraging of Mobile DevicesrdquoIEEE Communications Surveys ampTutorials vol PP no 99 pp 1ndash12 2011

[20] R Buyya C S Yeo S Venugopal J Broberg and I Brandic ldquoCloudcomputing and emerging IT platforms Vision hype and reality fordelivering computing as the 5th utilityrdquoFuture Generation ComputerSystems vol 25 no 6 pp 599ndash616 2009

[21] P Mell and T Grance ldquoThe NIST Definition of CloudComputing Recommendations of the National Institute of Standardsand Technologyrdquo 2011 [Online] Available httpcsrcnistgovpublicationsnistpubs800-145SP800-145pdf

[22] R Buyya J Broberg and A GoscinskiCloud computing Principlesand paradigms Wiley 2010

[23] M Hogan F Liu A Sokol and J Tong ldquoNIST Cloud ComputingStandards Roadmaprdquo Jul 2011 [Online] Available httpwwwnistgovitlclouduploadNISTSP-500-291Jul5Apdf

[24] G Huerta-Canepa and D Lee ldquoA Virtual Cloud ComputingProviderfor Mobile Devicesrdquo in Proc ACM MCSrsquo10 Social Networks andBeyond San Francisco USA 2010 pp 1ndash5

[25] E Cuervo A Balasubramanian D Cho A Wolman S SaroiuR Chandra and P Bahl ldquoMAUI making smartphones last longer withcode offloadrdquo inProc ACM MobiSysrsquo10 San Francisco CA USA2010 pp 49ndash62

[26] M Satyanarayanan P Bahl R Caceres and N Davies ldquoThe Case forVM-Based Cloudlets in Mobile ComputingrdquoIEEE Pervasive Comput-ing vol 8 no 4 pp 14ndash23 2009

[27] T Verbelen P Simoens F De Turck and B Dhoedt ldquoAIOLOSMiddleware for improving mobile application performance throughcyber foragingrdquo Journal of Systems and Software vol 85 no 11pp 2629 ndash 2639 2012

[28] S-H Hung C-S Shih J-P Shieh C-P Lee and Y-H HuangldquoExecuting mobile applications on the cloud Framework andissuesrdquoComputers and Mathematics with Applications vol 63 no 2 pp 573ndash587 2011

[29] R Kemp N Palmer T Kielmann F Seinstra N Drost J Maassenand H Bal ldquoeyeDentify Multimedia Cyber Foraging from a Smart-phonerdquo inProc ISMrsquo09 San Diego California USA 2009 pp 392ndash399

[30] S Kosta A Aucinas P Hui R Mortier and X Zhang ldquoThinkAirDynamic resource allocation and parallel execution in the cloud formobile code offloadingrdquoProc IEEE INFOCOM 12 pp 945ndash 9532012

[31] Y Guo L Zhang J Kong J Sun T Feng and X Chen ldquoJupitertransparent augmentation of smartphone capabilities through cloudcomputingrdquo in Proc ACM MobiHeld rsquo11 Cascais Portugal 2011p 2

[32] AManjunatha ARanabahu ASheth and KThirunarayan ldquoPower ofClouds In Your Pocket An Efficient Approach for Cloud MobileHybrid Application Developmentrdquo inProc CloudComrsquo10 IndianaUSA 2010 pp 496ndash503

[33] X W Zhang A Kunjithapatham S Jeong and S Gibbs ldquoTowards anElastic Application Model for Augmenting the Computing Capabilities

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 29: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 29

of Mobile Devices with Cloud ComputingrdquoMobile Networks ampApplications vol 16 no 3 pp 270ndash284 2011

[34] B G Chun S Ihm P Maniatis M Naik and A Patti ldquoCloneCloudElastic execution between mobile device and cloudrdquo inProc ACMEuroSysrsquo11 Salzburg Austria 2011 pp 301ndash314

[35] B G Chun and P Maniatis ldquoAugmented smartphone applicationsthrough clone cloud executionrdquo inProc HotOSrsquo09 Monte VeritaSwitzerland 2009 pp 8ndash14

[36] V March Y Gu E Leonardi G Goh M Kirchberg and BS LeeldquomicroCloud Towards a New Paradigm of Rich Mobile ApplicationsrdquoinProc IEEE MobWISrsquo11 Ontario Canada 2011

[37] X Luo ldquoFrom Augmented Reality to Augmented Computing a Lookat Cloud-Mobile ConvergencerdquoProc IEEE ISUVR09 pp 29ndash32 2009

[38] E Badidi and I Taleb ldquoTowards a cloud-based framework for contextmanagementrdquo inProc IEEE IITrsquo11 Abu Dhabi United Arab Emirates2011 pp 35ndash40

[39] Q Liu X Jian J Hu H Zhao and S Zhang ldquoAn Optimized Solutionfor Mobile Environment Using Mobile Cloud Computingrdquo inProcWiComrsquo09 Beijing China 2009 pp 1ndash5

[40] K Kumar and Y H Lu ldquoCloud Computing for Mobile UsersCanOffloading Computation Save EnergyrdquoIEEE Computer vol 43 no 4pp 51ndash56 2010

[41] B-G Chun and P Maniatis ldquoDynamically PartitioningApplicationsbetween Weak Devices and Cloudsrdquo in1st ACM Workshop MCSrsquo10Social Networks and Beyond San Francisco USA 2010 pp 1ndash5

[42] Y Lu S P Li and H F Shen ldquoVirtualized Screen A Third Elementfor Cloud-Mobile ConvergencerdquoIEEE Multimedia vol 18 no 2 pp4ndash11 2011

[43] RKemp N Palmer T Kielmann and H Bal ldquoCuckoo a ComputationOffloading Framework for Smartphonesrdquo inProc MobiCASErsquo10Santa Clara CA USA 2010

[44] R Kemp N Palmer T Kielmann and H Bal ldquoThe Smartphone andthe Cloud Power to the Userrdquo inProc MobiCASE rsquo12 CaliforniaUSA 2012 pp 342ndash348

[45] S Abolfazli Z Sanaei M Shiraz and A Gani ldquoMOMCCMarket-oriented architecture for Mobile Cloud Computing based on ServiceOriented Architecturerdquo inProc IEEE MobiCCrsquo12 Beijing China2012 pp 8ndash 13

[46] S Sanaei Z Abolfazli M Shiraz and A Gani ldquoSAMI Service-Based Arbitrated Multi-Tier Infrastructure Model for Mobile CloudComputingrdquo inProc IEEE MobiCCrsquo12 Beijing China 2012 pp 14ndash19

[47] R K K Ma and C L Wang ldquoLightweight Application-Level TaskMigration for Mobile Cloud Computingrdquo inProc IEEE AINArsquo122012 pp 550ndash557

[48] Y Gu V March and B S Lee ldquoGMoCA Green mobile cloudapplicationsrdquo inProc IEEE GREENSrsquo12 Zurich Switzerland Jun2012 pp 15ndash20

[49] I Giurgiu O Riva D Julic I Krivulev and G Alonso ldquoCalling theCloud Enabling Mobile Phones as Interfaces to Cloud ApplicationsrdquoProc ACM Middleware09 vol 5896 pp 83ndash102 2009

[50] Z Ou H Zhuang J K Nurminen A Yla-Jaaski and P HuildquoExploiting hardware heterogeneity within the same instance type ofAmazon EC2rdquo in Proc USENIX HotCloudrsquo12 Boston USA Jun2012 p 4

[51] Z Sanaei S Abolfazli A Gani and R K Buyya ldquoHeterogeneityin Mobile Cloud Computing Taxonomy and Open ChallengesrdquoIEEECommunications Surveys amp Tutorials 2013

[52] K Salah and R Boutaba ldquoEstimating service response time for elasticcloud applicationsrdquo inProc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 12ndash16

[53] J W Smith A Khajeh-Hosseini J S Ward and I SommervilleldquoCloudMonitor Profiling Power Usagerdquo inProc IEEE CLOUD rsquo12Jun 2012 pp 947ndash948

[54] M Di Francesco ldquoEnergy consumption of remote desktopaccesson mobile devices An experimental studyrdquo inProc IEEE CLOUD-NETrsquo12 Paris Nov 2012 pp 105ndash110

[55] J Heide F H P Fitzek M V Pedersen and M Katz ldquoGreen mobileclouds Network coding and user cooperation for improved energyefficiencyrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 111ndash118

[56] M Shiraz A Gani R Hafeez Khokhar and R Buyya ldquoA Reviewon Distributed Application Processing Frameworks in SmartMobileDevices for Mobile Cloud ComputingrdquoIEEE Communications Surveysamp Tutorials Nov 2012

[57] N Fernando S W Loke and W Rahayu ldquoMobile cloud computingA surveyrdquo Future Generation Computer Systems vol 29 no 2 pp84ndash106 2012

[58] H T Dinh C Lee D Niyato and P Wang ldquoA survey of mobilecloud computing architecture applications and approachesrdquo WirelessCommunications and Mobile Computing 2011

[59] I Foster C Kesselman and S Tuecke ldquoThe anatomy of the gridEnabling scalable virtual organizationsrdquoInternational journal of highperformance computing applications vol 15 no 3 pp 200ndash222 2001

[60] Z Li C Wang and R Xu ldquoComputation offloading to saveenergyon handheld devices a partition schemerdquo inProc ACM CASES rsquo01Atlanta Georgia USA 2001 pp 238ndash246

[61] Z Li and R Xu ldquoEnergy impact of secure computation on ahandhelddevicerdquo in Proc IEEE WWC-5 rsquo02 Austin Texas USA 2002 pp109ndash117

[62] A Athan and D Duchamp ldquoAgent-mediated message passing forconstrained environmentsrdquo inProc USENIX MLCS rsquo93 CambridgeMassachusetts 1993 pp 103ndash107

[63] A Bakre and B R Badrinath ldquoM-RPC A remote procedurecallservice for mobile clientsrdquo inProc ACM MobiComrsquo95 BerkeleyCalifornia 1995 pp 97ndash110

[64] A Rudenko P Reiher G J Popek and G H Kuenning ldquoThe remoteprocessing framework for portable computer power savingrdquoin ProcACM SAC rsquo99 San Antonio Texas USA 1999 pp 365ndash372

[65] J M Wrabetz D D Mason Jr and M P Gooderum ldquoIntegratedremote execution system for a heterogenous computer network envi-ronmentrdquo Patent US Patent 5442791 1995

[66] K Kumar J Liu Y H Lu and B Bhargava ldquoA Survey ofComputation Offloading for Mobile SystemsrdquoMobile Networks andApplications pp 1ndash12 2012

[67] K Bent (2012 May) Obama Government Agencies Have OneYear To Deploy Smartphone-Friendly Services [Online] Availablehttpwwwcrncomnewsmobility240000931obama-government-agencies-have-one-year-to-deploy-smartphone-friendly-serviceshtmcid=rssFeed

[68] T Khalifa K Naik and A Nayak ldquoA Survey of CommunicationProtocols for Automatic Meter Reading ApplicationsrdquoIEEE Commu-nications Surveys amp Tutorials vol 13 no 2 pp 168ndash182 2011

[69] M Satyanarayanan S of Computer Science C Mellonand Univer-sity ldquoMobile Computing the Next Decaderdquo inProc ACM MCS rsquo10San Francisco USA 2010

[70] J Park and B Choi ldquoAutomated Memory Leakage Detection inAndroid Based SystemsrdquoInternational Journal of Control and Au-tomation vol 5 issue 2 2012

[71] G Xu and A Rountev ldquoPrecise memory leak detection forjavasoftware using container profilingrdquo inProc ACMIEEE ICSE rsquo08Leipzig Germany May 2008 pp 151ndash160

[72] M Satyanarayanan ldquoAvoiding Dead BatteriesrdquoIEEE Pervasive Com-puting vol 4 no 1 pp 2ndash3 2005

[73] A P Miettinen and J K Nurminen ldquoEnergy efficiency ofmobileclients in cloud computingrdquo inProc USENIX HotCloudrsquo10 BostonMA 2010 pp 4ndash11

[74] Y Neuvo ldquoCellular phones as embedded systemsrdquoDigest of TechnicalPapers IEEE ISSCC rsquo04 vol 47 no 1 pp 32ndash37 2004

[75] S Robinson ldquoCellphone Energy Gap Desperately Seeking Solutionsrdquop 28 2009 [Online] Available httpstrategyanalyticscomdefaultaspxmod=reportabstractviewerampa0=4645

[76] D Ben B Ma L Liu Z Xia W Zhang and F Liu ldquoUnusual burnswith combined injuries caused by mobile phone explosion watch outfor the ldquomini-bombrdquordquo Journal of Burn Care and Research vol 30no 6 p 1048 2009

[77] Cellular-News (2007) Chinese Man Killed by ExplodingMobilePhone [Online] Available httpwwwcellular-newscomstory24754php

[78] J Flinn and M Satyanarayanan ldquoEnergy-aware adaptation for mobileapplicationsrdquo inProc ACM SOSPrsquo 99 vol 33 1999 pp 48ndash63

[79] T Starner D Kirsch and S Assefa ldquoThe Locust SwarmAnenvironmentally-powered networkless location and messaging sys-temrdquo in Proc IEEE ISWCrsquo 97 Boston MA USA 1997 pp 169ndash170

[80] S RAvro ldquoWireless Electricity is Real and Can Changethe Worldrdquo2009 [Online] Available httpwwwconsumerenergyreportcom20090115wireless-electricity-is-real-and-can-change-the-world

[81] W F Pickard and D Abbott ldquoAddressing the Intermittency ChallengeMassive Energy Storage in a Sustainable Future [Scanning the Issue]rdquoProceedings of the IEEE vol 100 no 2 pp 317ndash321 2012

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 30: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 30

[82] A P Bianzino C Chaudet D Rossi and J-L RougierldquoA Surveyof Green Networking ResearchrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 3ndash20 2012

[83] N Vallina-Rodriguez and J Crowcroft ldquoEnergy Management Tech-niques in Modern Mobile HandsetsrdquoIEEE Communications Surveysamp Tutorials vol 15 no 1 pp 179ndash198 2013

[84] K Jackson (2009) Missouri University Researchers CreateSmaller and More Efficient Nuclear Battery [Online]Available httpmunewsmissouriedunews-releases20091007-mu-researchers-create-smaller-and-more-efficient-nuclear-battery

[85] J F Gantz C Chute A Manfrediz S Minton D Reinsel W Schlicht-ing and A Toncheva ldquoThe Diverse and Exploding Digital UniverseAn Updated Forecast of Worldwide Information Growth Through2011rdquo White Paper 2008

[86] X F Guo J J Shen and S M Wu ldquoOn Workflow Engine Based onService-Oriented ArchitecturerdquoProc IEEE ISISE rsquo08 pp 129ndash1322008

[87] S Ortiz ldquoBringing 3D to the Small ScreenrdquoIEEE Computer vol 44no 10 pp 11ndash13 2011

[88] T Capin K Pulli and T Akenine-Moller ldquoThe State ofthe Art in Mo-bile Graphics ResearchrdquoIEEE Computer Graphics and Applicationsvol 28 no 4 pp 74ndash84 2008

[89] Prosper Mobile Insights ldquoSmartphoneTablet User Surveyrdquo 2011[Online] Available httpprospermobileinsightscomDefaultaspxpg=19

[90] M La Polla F Martinelli and D Sgandurra ldquoA Survey on Security forMobile DevicesrdquoIEEE Communications Surveys amp Tutorials 2012

[91] Z Xiao and Y Xiao ldquoSecurity and Privacy in Cloud ComputingrdquoIEEE Communications Surveys amp Tutorials vol 15 no 2 pp 843ndash859 2013

[92] C Cachin and M Schunter ldquoA cloud you can trustrdquoIEEE Spectrumvol 48 no 12 pp 28ndash51 Dec 2011

[93] NVidia (2010) The Benefits of Multiple CPU Cores in Mobile Devices[Online] Available httpwwwnvidiacomcontentPDFtegra whitepapersBenefits-of-Multi-core-CPUs-in-Mobile-DevicesVer12pdf

[94] ARM (2009) The ARM Cortex-A9 Processors Version 20 WhitePaper [Online] Available httparmcomfilespdfARMCortexA-9Processorspdf

[95] M Altamimi R Palit K Naik and A Nayak ldquoEnergy-as-a-Service(EaaS) On the Efficacy of Multimedia Cloud Computing to SaveSmartphone Energyrdquo inProc IEEE CLOUD rsquo12 Honolulu HawaiiUSA Jun 2012 pp 764ndash771

[96] K Fekete K Csorba B Forstner M Feher and T VajkldquoEnergy-efficient computation offloading model for mobile phone environmentrdquoin Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 95ndash99

[97] S Park and B-S Moon ldquoDesign and Implementation of iSCSI-based Remote Storage System for Mobile AppliancerdquoProc IEEEHEALTHCOM rsquo05 pp 236ndash240 2003

[98] G Lawton ldquoCloud Streaming Brings Video to Mobile Devicesrdquo IEEEComputer vol 45 no 2 pp 14ndash16 2012

[99] B Seshasayee R Nathuji and K Schwan ldquoEnergy-aware mobile ser-vice overlays Cooperative dynamic power management in distributedmobile systemsrdquo inProc IEEE ICACrsquo07 Jacksonville Florida USA2007 p 6

[100] Y J Hong K Kumar and Y H Lu ldquoEnergy efficient content-basedimage retrieval for mobile systemsrdquo inProc IEEE ISCAS rsquo09 TaipeiTaiwan 2009 pp 1673ndash1676

[101] B Rajesh Krishna ldquoPowerful change part 2 reducing the powerdemands of mobile devicesrdquoIEEE Pervasive Computing vol 3 no 2pp 71ndash73 2004

[102] A F Murarasu and T Magedanz ldquoMobile middleware solution forautomatic reconfiguration of applicationsrdquo inProc ITNG rsquo09 LasVegas Nevada USA 2009 pp 1049ndash1055

[103] E Abebe and C Ryan ldquoAdaptive application offloadingusing dis-tributed abstract class graphs in mobile environmentsrdquoJournal ofSystems and Software vol 85 no 12 pp 2755ndash2769 2012

[104] M Salehie and L Tahvildari ldquoSelf-adaptive software Landscape andresearch challengesrdquoACM Transactions on Autonomous and AdaptiveSystems (TAAS) vol 4 no 2 p 14 2009

[105] G Huerta-Canepa and D Lee ldquoAn adaptable application offloadingscheme based on application behaviorrdquo inProc IEEE AINAW rsquo08GinoWan Okinawa Japan 2008 pp 387ndash392

[106] F SchmidtThe SCSI bus and IDE interface protocols applicationsand programming Addison-Wesley Longman Publishing Co Inc1997

[107] Y Lu and D H C Du ldquoPerformance study of iSCSI-basedstoragesubsystemsrdquoIEEE Communications Magazine vol 41 no 8 pp 76ndash82 2003

[108] J-H Kim B-S Moon and M-S Park ldquoMiSC A New AvailabilityRemote Storage System for Mobile Appliancerdquo inICN 2005 serLNCS 3421 P Lorenz and P Dini Eds Springer 2005 pp 504ndash 520

[109] M Ok D Kim and M Park ldquoUbiqStor Server and Proxy for RemoteStorage of Mobile DevicesrdquoEmerging Directions in Embedded andUbiquitous Computing pp 22ndash31 2006

[110] M H OK D Kim and M Park ldquoUbiqStor A remote storage servicefor mobile devicesrdquoParallel and Distributed Computing Applicationsand Technologies pp 53ndash74 2005

[111] D Kim M Ok and M Park ldquoAn Intermediate Target for Quick-Relayof Remote Storage to Mobile Devicesrdquo inComputational Science andIts Applications - ICCSA 2005 ser Lecture Notes in Computer ScienceSpringer Berlin Heidelberg 2005 vol 3481 pp 91ndash100

[112] W Zheng P Xu X Huang and N Wu ldquoDesign a cloud storageplatform for pervasive computing environmentsrdquoCluster Computingvol 13 no 2 pp 141ndash151 2010

[113] U Kremer J Hicks and J Rehg ldquoA compilation framework forpower and energy management on mobile computersrdquoLanguages andCompilers for Parallel Computing pp 115ndash131 2003

[114] S Gurun and C Krintz ldquoAddressing the energy crisis in mobilecomputing with developing power aware softwarerdquo vol 8 no64MB 2003 [Online] Available httpwwwcsucsbeduresearchtech reportsreports2003-15ps

[115] X Ma Y Cui and I Stojmenovic ldquoEnergy Efficiency onLocationBased Applications in Mobile Cloud Computing A SurveyrdquoProcediaComputer Science vol 10 pp 577ndash584 2012

[116] G P Perrucci F H P Fitzek and J Widmer ldquoSurvey on EnergyConsumption Entities on the Smartphone Platformrdquo inProc IEEEVTC Spring rsquo11 Budapest Hungary 2011 pp 1ndash6

[117] E D Lara D S Wallach and W Zwaenepoel ldquoPuppeteerComponent-based adaptation for mobile computingrdquo inProc USENIXUSITS 2001 pp 14ndash25

[118] J H Christensen ldquoUsing RESTful web-services and cloud computingto create next generation mobile applicationsrdquo inProc ACM OOPSLArsquo09 Orlando Florida USA 2009 pp 627ndash634

[119] M Shiraz S Abolfazli Z Sanaei and A Gani ldquoA study on virtualmachine deployment for application outsourcing in mobile cloud com-putingrdquo The Journal of Supercomputing vol 63 no 3 pp 946ndash964Dec 2012

[120] Z Sanaei S Abolfazli A Gani and R H Khokhar ldquoTripod ofRequirements in Horizontal Heterogeneous Mobile Cloud Computingrdquoin Proc WSEAS CISCOrsquo12 2012

[121] H Liu H Jin C-Z Xu and X Liao ldquoPerformance and energymodeling for live migration of virtual machinesrdquoCluster Computingpp 1ndash16 2011

[122] B Gao L He L Liu K Li and S A Jarvis ldquoFrom Mobiles toClouds Developing Energy-Aware Offloading Strategies forWork-flowsrdquo in Proc ACMIEEE GRID rsquo12 Dubna 2012 pp 139ndash146

[123] W Zeng Y Zhao K Ou and W Song ldquoResearch on cloud storagearchitecture and key technologiesrdquo inProc ACM ICIS rsquo09 SeoulRepublic of Korea 2009 pp 1044ndash1048

[124] J Yang H Wang J Wang C Tan and D Yu ldquoProvable Data Pos-session of Resource-constrained Mobile Devices in Cloud ComputingrdquoJournal of Networks vol 6 no 7 pp 1033ndash1040 2011

[125] N Nasser A Hasswa and H Hassanein ldquoHandoffs in fourth gen-eration heterogeneous networksrdquoIEEE Communications Magazinevol 44 no 10 pp 96ndash103 2006

[126] ldquoMobile Cloud - Converged Fix Wireless and Mobile Infrastructurerdquo2013 [Online] Available httpsieneoncominfrashtm

[127] ldquoConverged Highly Integrated Network Access (CHINA)rdquo 2013[Online] Available httpsieneoncomnetworkhtm

[128] R Kuhne G Huitema and G Carle ldquoCharging and Billing in ModernCommunications Networks A Comprehensive Survey of the State ofthe Art and Future RequirementsrdquoIEEE Communications Surveys ampTutorials vol 14 no 1 pp 170ndash192 2012

[129] (2013) Panda Global Business Protection on-premise securitysolutions [Online] Available httpwwwpandasecuritycomusaenterprisesecurity-solutions-on-premise

[130] S Kamara and K Lauter ldquoCryptographic cloud storagerdquo FinancialCryptography and Data Security pp 136ndash149 2010

[131] C Wang Q Wang K Ren and W Lou ldquoEnsuring data storagesecurity in cloud computingrdquo inProc IEEE IWQoS rsquo09 CharlestonSouth Carolina USA 2009 pp 1ndash9

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 31: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 31

[132] T Mather S Kumaraswamy and S LatifCloud security and privacyan enterprise perspective on risks and compliance OrsquoReilly MediaIncorporated 2009

[133] J Clark (2013 Feb) Microsoft secure Azure Storagegoes downWordwide [Online] Available httpwwwtheregistercouk20130222azureproblem that should never happenever

[134] A Clark Christopher and Fraser Keir and Hand Steven and HansenJacob Gorm and Jul Eric and Limpach Christian and Pratt Ian andWarfield ldquoLive migration of virtual machinesrdquo inProc USENIX NSDIrsquo05 Boston MA USA 2005 pp 273ndash286

[135] K Owens Securing Virtual Compute Infras-tructure in the Cloud Whitepaper [Online]Available httpwwwsavviscomen-usinfocenterdocumentshos-whitepaper-securingvirutalcomputeinfrastructureinthecloudpdf

[136] S Garriss R Caceres S Berger R Sailer L van Doorn andX Zhang ldquoTrustworthy and personalized computing on public kiosksrdquoin Proc ACM MobiSys rsquo08 Breckenridge CO USA 2008 pp 199ndash210

[137] N D Lane E Miluzzo L Hong D Peebles T Choudhury and A TCampbell ldquoA survey of mobile phone sensingrdquoIEEE CommunicationsMagazine vol 48 no 9 pp 140ndash150 2010

[138] P Lukowicz S Pentland and A Ferscha ldquoFrom Context Awarenessto Socially Aware ComputingrdquoPervasive Computing vol 11 no 1pp 32ndash41 Jan 2012

[139] P Makris D N Skoutas and C Skianis ldquoA Survey on Context-AwareMobile and Wireless Networking On Networking and Computing En-vironmentsrsquo IntegrationrdquoIEEE Communications Surveys amp Tutorialsvol 15 no 1 pp 362ndash386 2013

[140] S Shen and B Blau (2012 Aug) Market Trends MobileAppStores Worldwide 2012 [Online] Available httpwwwgartnercomid=2126015

[141] M ANIA (2012) Beware Free Apps Might ProveCostly [Online] Available httptheinstituteieeeorgtechnology-focustechnology-topicbeware-free-apps-might-prove-costly

[142] W Enck M Ongtang and P McDaniel ldquoOn lightweight mobile phoneapplication certificationrdquo inProc IEEE CCSrsquo11 Milan Italy 2011 pp235ndash245

[143] M Rahimi N Venkatasubramanian S Mehrotra and VVasilakosldquoMAPCloud mobile applications on an elastic and scalable 2-tier cloudarchitecturerdquo inProc IEEEACM UCCrsquo12 Chicago Illionis USA2012

[144] H Kim and M Parashar ldquoCometCloud An Autonomic Cloud EnginerdquoCloud Computing Principles and Paradigms pp 275ndash297 2011

[145] J Broberg R Buyya and Z Tari ldquoMetaCDN Harnessing StorageClouds for high performance content deliveryrdquoJournal of Networkand Computer Applications vol 32 no 5 pp 1012ndash1022 2009

[146] K Hariharan ldquoBest Practices Extending EnterpriseApplicationsto Mobile DevicesrdquoThe Architecture Journal Microsoft ArchitectureCenter vol 14 2008 [Online] Available httpmsdnmicrosoftcomen-uslibrarybb985493aspx

[147] A H Ranabahu E M Maximilien A P Sheth and K ThirunarayanldquoA domain specific language for enterprise grade cloud-mobile hy-brid applicationsrdquo inProc ACM co-located workshops on DSMrsquo11TMCrsquo11 AGERErsquo11 AOOPESrsquo11 NEATrsquo11 amp VMILrsquo11 PortlandOregon USA 2011 pp 77ndash84

[148] A van Deursen P Klint and J Visser ldquoDomain-Specific LanguagesAn Annotated BibliographyrdquoSIGPLAN Notices vol 35 no 6 pp26ndash36 2000

[149] P Stuedi I Mohomed and D Terry ldquoWhereStore Location-baseddata storage for mobile devices interacting with the cloudrdquo in ProcACM MCSrsquo10 San Francisco California USA 2010 p 1

[150] H Mao N Xiao W Shi and Y Lu ldquoWukong A cloud-oriented fileservice for mobile Internet devicesrdquoJournal of Parallel and DistributedComputing 2011

[151] N Tolia D G Andersen and M Satyanarayanan ldquoQuantifyinginteractive user experience on thin clientsrdquoIEEE Computer vol 39no 3 pp 46ndash52 2006

[152] E E Marinelli ldquoHyrax cloud computing on mobile devices usingMapReducerdquo Master Thesis Computer Science DeptartmentCarnegieMellon University 2009

[153] C Mei D Taylor C Wang A Chandra and J WeissmanldquoMobilizingthe Cloud Enabling Multi-User Mobile Outsourcing in the CloudrdquoDepartment of Computer Science and Engineering University ofMinnesota Tech Rep TR 11-029 2011

[154] C Mei D Taylor C Wang A Chandra and J WeissmanldquoSharing-Aware Cloud-Based Mobile Outsourcingrdquo inProc IEEE CLOUD rsquo12Hawaii USA Jun 2012 pp 408ndash415

[155] M Guirguis R Ogden Z Song S Thapa and Q Gu ldquoCanYou HelpMe Run These Code Segments on Your Mobile Devicerdquo inProc IEEEGLOBECOM rsquo11 Houston Texas USA Dec 2011 pp 1ndash5

[156] T White Hadoop The definitive guide OrsquoReilly Media 2012[157] J Dean and S Ghemawat ldquoMapReduce Simplified Data Processing

on Large ClustersrdquoCommunications of the ACM vol 51 no 1 pp107ndash113 2004

[158] T Soyata R Muraleedharan C Funai M Kwon and W HeinzelmanldquoCloud-Vision Real-time face recognition using a mobile-cloudlet-cloud acceleration architecturerdquo inProc IEEE ISCC rsquo12 CappadociaTurkey Jul 2012 pp 59ndash66

[159] P Bahl R Y Han L E Li and M Satyanarayanan ldquoAdvancing thestate of mobile cloud computingrdquo inProc ACM MCS rsquo12 ser MCSrsquo12 Helsinki Finland 2012 pp 21ndash28

[160] Gartner (2012 Feb) Worldwide Smartphone Sales Soared in FourthQuarter of 2011 With 47 Percent Growth [Online] Available httpwwwgartnercomitpagejspid=1924314

[161] H Viswanathan B Chen and D Pompili ldquoResearch Challenges inComputation Communication and Context awareness for UbiquitousHealthcarerdquoIEEE Communications Magazine vol 50 no 5 p 922012

[162] E Koukoumidis M Martonosi and L-S Peh ldquoLeveraging Smart-phone Cameras for Collaborative Road AdvisoriesrdquoIEEE Transactionson Mobile Computing vol 11 no 5 pp 707ndash723 May 2012

[163] M Kranz A Moller N Hammerla S Diewald T Plotz P Olivierand L Roalter ldquoThe mobile fitness coach Towards individualized skillassessment using personalized mobile devicesrdquoPervasive and MobileComputing vol 9 no 2 pp 203ndash215 2012

[164] P Bellavista A Corradi and C Giannelli ldquoA unifying perspective oncontext-aware evaluation and management of heterogeneouswirelessconnectivityrdquo IEEE Communications Surveys amp Tutorials vol 13no 3 pp 337ndash357 2011

[165] American National Standard Dictionary of Information Technology[Online] Available httpincitsorgANSDITp2htm[June-19-2012]

[166] X Yan Y Ahmet Sekercioglu and S Narayanan ldquoA survey of verticalhandover decision algorithms in Fourth Generation heterogeneouswireless networksrdquoComputer Networks vol 54 no 11 pp 1848ndash1863 2010

[167] S Sakr A Liu D M Batista and M Alomari ldquoA surveyof largescale data management approaches in cloud environmentsrdquoIEEECommunications Surveys amp Tutorials vol 13 no 3 pp 311ndash3362011

[168] H Lagar-Cavilla N Tolia E De Lara M Satyanarayanan andD OrsquoHallaron ldquoInteractive resource-intensive applications made easyrdquoin Proc MIDDLEWARE Springer 2007 pp 143ndash163

[169] R Bifulco and R Canonico ldquoAnalysis of the handover procedure inFollow-Me Cloudrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012pp 185ndash187

[170] D Johansson and K Andersson ldquoWeb-based adaptive applicationmobilityrdquo in Proc IEEE CLOUDNETrsquo12 Paris Nov 2012 pp 87ndash94

[171] B C Guevara Marisabel and Lubin Benjamin and Lee ldquoNavigatingHeterogeneous Processors with Market Mechanismsrdquo inProc IEEEHPCArsquo13 Shenzhen China 2013

[172] R Buyya R N Calheiros and X Li ldquoSpecial Section on AutonomicCloud Computing Technologies Services and Applicationsrdquo Concur-rency and Computation Practice and Experience vol 24 no 9 pp935ndash937 2012

[173] P Yu X Ma J Cao and J Lu ldquoApplication mobility inpervasivecomputing A surveyrdquoPervasive and Mobile Computing vol 9 no 1pp 2ndash17 Feb 2013

[174] G Blair M Paolucci P Grace and N Georgantas ldquoInteroperabilityin Complex Distributed Systemsrdquo inFormal Methods for EternalNetworked Software Systems ser Lecture Notes in Computer ScienceM Bernardo and V A Issarny Eds Springer Berlin Heidelberg2011 vol 6659 pp 1ndash26

[175] ARanabahu and ASheth ldquoSemantics Centric Solutions for Applicationand Data Portability in Cloud Computingrdquo inProc IEEE Cloud-Comrsquo10 Indianapolis USA 2010 pp 234ndash241

[176] L Youseff R Wolski B Gorda and C Krintz ldquoParavirtualizationfor HPC Systemsrdquo inFrontiers of High Performance Computing andNetworking-ISPA 2006 Workshops ser Lecture Notes in ComputerScience G Min B Martino L Yang M Guo and G Ruenger Eds2006 vol 4331 pp 474ndash486

[177] Q Wang C Wang J Li K Ren and W Lou ldquoEnabling public ver-ifiability and data dynamics for storage security in cloud computingrdquoProc Springer ESORICS rsquo09 pp 355ndash370 2009

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya
Page 32: Cloud-based augmentation for mobile devices: Motivation, Taxonomy, and Open Challenges

IEEE COMMUNICATIONS SURVEYS amp TUTORIALS ACCEPTED FOR PUBLICATION 32

[178] W Wang Z Li R Owens and B Bhargava ldquoSecure and efficientaccess to outsourced datardquo inProc ACM CCSrsquo 09 Chicago IL USA2009 pp 55ndash66

[179] M Mowbray and S Pearson ldquoA client-based privacy manager forcloud computingrdquo inProc ACM COMSWARE rsquo09 Dublin Ireland2009 p 5

[180] S Ruj A Nayak and I Stojmenovic ldquoDACC Distributed accesscontrol in cloudsrdquo inProc IEEE TRUSTCOM rsquo 11 Changsha China2011 pp 91ndash98

[181] ldquoVirtual Machine Migration Comparision VMwareVsphere VSMicrosoft Hyper-Vrdquo p 36 2011 [Online] Available httpwwwvmwarecomfilespdfvmw-vmotion-verus-live-migrationpdf

[182] T Takahashi Kazushi and Sasada Koichi and Hirofuchi ldquoA Fast Vir-tual Machine Storage Migration Technique Using Data Deduplicationrdquoin Proc CLOUD COMPUTINGrsquo12 Nice France 2012 pp 57ndash64

[183] R Watanabe Hidenobu and Ohigashi Toshihiro and Kondo Tohru andNishimura Kouji and Aibara ldquoA Performance Improvement Methodfor the Global Live Migration of Virtual Machine with IP Mobilityrdquoin Proc IEEEIPSJ ICMU rsquo10 Seattle Washington USA 2010

[184] M Harney Eric and Goasguen Sebastien and Martin Jim and MurphyMike and Westall ldquoThe efficacy of live virtual machine migrations overthe internetrdquo inProc ACM VTDC rsquo07 Reno Nevada USA 2007

[185] G Raad Patrick and Colombo Giulio and Chi Dung Phung andSecci Stefano and Cianfrani Antonio and Gallard Pascal and PujolleldquoAchieving Sub-Second Downtimes in Internet-wide VirtualMachineLive Migrations in LISP Networksrdquo inProc IFIPIEEE Symposiumon Integrated Network Management Ghent Belgium 2013

Saeid Abolfazli is currently a PhD candidateresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation and part time lecturer in the Departmentof Computer Systems and Technology at the Uni-versity of Malaya Malaysia He received his MScin Information Systems in 2008 from India and BE(Software Engineering) in 2001 from Iran He hasbeen serving as CEO of Espanta Information Com-plex during 1999-2006 in Iran He also was part time

lecturer to the ministry of education and Khorasan Technical and VocationalOrganization between 2000 and 2006 He is a member of IEEE societyand IEEE CS Cloud Computing STC He has been serving as a reviewerfor several international conference and ISI journals of computer scienceHis main research interests include Mobile Cloud Computing lightweightprotocols and service oriented computing (SOC) Please write to him atabolfazlisgmailcom or abolfazliieeeorg For further information pleasevisit his cyberhome at wwwmobilecloudfamilycom

Zohreh Sanaei is currently a PhD candidate andresearch assistant in High Impact Research Project(Mobile Cloud Computing Device and Connectiv-ity) fully funded by Malaysian Ministry of HigherEducation in the Department of Computer Sys-tems and Technology at the University of MalayaMalaysia She received her MSc in InformationSystems in 2008 from India and BE (Software En-gineering) in 2001 from Iran She worked in 3MCDand EIC Iran as a network engineer and participatedin several wireless communication projects from

2001 till 2006 She has been working for more than 6 years as a part-time lec-turer for the ministry of social affairs Iran as a technicaland vocational trainerHer main research interests include mobile cloud computing distributedcomputing and ubiquitous computing She is a member of IEEEsociety andcan be corresponded via zsanaeimgmailcom or sanaeiieeeorg For furtherinformation please visit her cyberhome at wwwmobilecloudfamilycom

Ejaz Ahmed was born in Gandhian MansehraPakistan He did his BS (Computer Science) fromAllama Iqbal Open University Islamabad PakistanAfterward he completed his MS (Computer Sci-ence) from Mohammad Ali Jinnah University Is-lamabad in 2009 Currently he is pursuing hisPhD Candidature under Bright Spark Program atFaculty of Computer and Information TechnologyUniversity Malaya Kuala Lumpur Malaysia He isan active researcher in Mobile Cloud ComputingResearch Group at Faculty of Computer Science and

Information Technology University Malaya Kuala LumpurMalaysia Hisareas of interest include Seamless Application Execution Framework Designfor Mobile Cloud Computing Designing of Channel Assignment and RoutingAlgorithms for Cognitive Radio Networks

Dr Abdullah Gani is Associate Professor of Com-puter System and Technology at the Universityof Malaya Malaysia His academic qualificationswere obtained from UKrsquos universities - bachelorand master degrees from the University of Hulland PhD from the University of Sheffield He hasvast teaching experience due to having worked invarious educational institutions locally and abroad -schools teaching college ministry of education anduniversities His interest in research started in 1983when he was chosen to attend Scientific Research

course in RECSAM by the Ministry of Education Malaysia More than100 academic papers have been published in conferences and respectablejournals He actively supervises many students at all levelof study - BachelorMaster and PhD His interest of research includes self-organized systemsreinforcement learning and wireless-related networks He is now working onmobile cloud computing with High Impact Research Grant for the period of2011-2016

Dr Rajkumar Buyya is Professor of ComputerScience and Software Engineering Future Fellowof the Australian Research Council and Directorof the Cloud Computing and Distributed Systems(CLOUDS) Laboratory at the University of Mel-bourne Australia He is also serving as the foundingCEO of Manjrasoft a spin-off company of theUniversity commercializing its innovations in CloudComputing He has authored over 425 publicationsand four text books including ldquoMastering CloudComputingrdquo published by McGraw Hill and Else-

vierMorgan Kaufmann 2013 for Indian and international markets respec-tively He also edited several books including ldquoCloud Computing Principlesand Paradigmsrdquo (Wiley Press USA Feb 2011) He is one of the highly citedauthors in computer science and software engineering worldwide (h-index=70g-index=144 23000+ citations) Microsoft Academic Search Index rankedDr Buyya as the worldrsquos top author in distributed and parallel computingbetween 2007 and 2012 Recently ISI has identified him as a ldquoHighly CitedResearcherrdquo based on citations to his journal papers

Software technologies for Grid and Cloud computing developed underDr Buyyarsquos leadership have gained rapid acceptance and arein use atseveral academic institutions and commercial enterprisesin 40 countriesaround the world Dr Buyya has led the establishment and developmentof key community activities including serving as foundation Chair of theIEEE Technical Committee on Scalable Computing and five IEEEACMconferences These contributions and international research leadership of DrBuyya are recognized through the award of ldquo2009 IEEE Medal for Excellencein Scalable Computingrdquo from the IEEE Computer Society USAManjrasoftrsquosAneka Cloud technology developed under his leadership has received ldquo2010Asia Pacific Frost amp Sullivan New Product Innovation Awardrdquo and ldquo2011Telstra Innovation Challenge Peoplersquos Choice Awardrdquo He is currently servingas the foundation Editor-in-Chief (EiC) of IEEE Transactions on CloudComputing For further information on Dr Buyya please visit his cyberhomewwwbuyyacom

  • Introduction
  • Mobile Computation Augmentation
    • Definition
    • Motivation
      • Processing Power
      • Energy Resources
      • Local Storage
      • Visualization Capabilities
      • Security Privacy and Data Safety
        • Mobile Augmentation Types Taxonomy
          • Impacts of CMA on Mobile Computing
            • Advantages
              • Empowered Processing
              • Prolonged Battery
              • Expanded Storage
              • Increased Data Safety
              • Ubiquitous Data Access and Content Sharing
              • Protected Offloaded Content
              • Enriched User Interface
              • Enhanced Application Generation
                • Disadvantages
                  • Dependency to High Performance Networking Infrastructure
                  • Excessive Communication Overhead and Traffic
                  • Unauthorized Access to Offloaded Data
                  • Application Development Complexity
                  • Paid Infrastructures
                  • Inconsistent Cloud Policies and Restrictions
                  • Service Negotiation and Control
                      • Taxonomy of Cloud-based Computing Resources
                        • Distant Immobile Clouds
                        • Proximate Immobile Computing Entities
                        • Proximate Mobile Computing Entities
                        • Hybrid (Converged Proximate and Distant Computing Entities)
                          • The State-of-the-art CMA Approaches Taxonomy
                            • Distant Fixed
                            • Proximate Fixed
                            • Proximate Mobile
                            • Hybrid
                              • CMA Prospectives
                                • CMA Decision Making Factors
                                  • Mobile Devices
                                  • Contents
                                  • Augmentation Environment
                                  • User Preferences and Requirements
                                  • Cloud Servers
                                    • Performance Limitation Factors
                                      • Heterogeneity
                                      • Data Volume
                                      • Round-Trip Latency
                                      • Context Management and Processing
                                      • Service Execution and Delivery
                                        • CMA Feasibility
                                          • Open Challenges
                                            • Reference Architecture for CMA Development
                                            • Autonomic CMA
                                            • Application Mobility Provisioning
                                            • Computing and Temporal Cost of Mobile Distributed Execution
                                            • Seamless Communication
                                            • Multipoint Data Bridging
                                            • Distributed Content Management
                                            • SeamlessLightweight CMA
                                            • Security in Mobile Cloud
                                            • Live Virtual Machine Migration
                                              • Conclusions
                                              • References
                                              • Biographies
                                                • Saeid Abolfazli
                                                • Zohreh Sanaei
                                                • Ejaz Ahmed
                                                • Dr Abdullah Gani
                                                • Dr Rajkumar Buyya