9 – Sensor Networks - Politecnico di Milanohome.deib.polimi.it/capone/rmd/9-Reti di...

Politecnico di MilanoFacoltà di Ingegneria dell’Informazione

9 – Sensor Networks

Reti Mobili DistribuiteProf. Antonio Capone

A. Capone: Reti mobili distribuite 2

Acknowlegments

This class notes are mostly based on the teaching material of:

Prof. Ian Akyildiz (Georgia Institute of Technology)Prof. Eylem Ekici (Ohio State University at Columbus)


Sensor Network Sensor Network ArchitectureArchitecture

Internet, Internet, Satellite, Satellite, UAVUAV

Sink

Sink

TaskManager

I.F. Akyildiz, W. Su, Y. Sankarasubramaniam, E. Cayirci, “Wireless Sensor Networks: A Survey”, Computer Networks (Elsevier) Journal, March 2002.


Characteristics of Characteristics of WSNsWSNs

Very large number of nodes, often in the order of thousandsVery large number of nodes, often in the order of thousandsNodes need to be close to each otherNodes need to be close to each otherDensities as high as 20 nodes/m3Densities as high as 20 nodes/m3Asymmetric flow of information, from sensor nodes to sinkAsymmetric flow of information, from sensor nodes to sinkCommunications are triggered by queries or eventsCommunications are triggered by queries or eventsLimited amount of energy (in many applications it is Limited amount of energy (in many applications it is impossible to replace or recharge)impossible to replace or recharge)Mostly static topologyMostly static topologyLow cost*, size, and weight per nodeLow cost*, size, and weight per nodeProne to failuresProne to failuresMore use of broadcast communications instead of pointMore use of broadcast communications instead of point--toto--pointpointNodes do not have a global ID such as an IP addressNodes do not have a global ID such as an IP addressThe security, both on physical and communication level, is The security, both on physical and communication level, is more limited than in classical wireless networks more limited than in classical wireless networks


Differences from AdDifferences from Ad--hoc hoc NetworksNetworksNumber of sensor nodes can be several orders Number of sensor nodes can be several orders of magnitude higherof magnitude higherSensor nodes are densely deployed and are Sensor nodes are densely deployed and are prone to failuresprone to failuresThe topology of a sensor network may change The topology of a sensor network may change frequently due to node failure and node frequently due to node failure and node mobilitymobilitySensor nodes are limited in power, Sensor nodes are limited in power, computational capacities, and memorycomputational capacities, and memoryMay not have global ID like IP addressMay not have global ID like IP addressNeed tight integration with sensing tasksNeed tight integration with sensing tasks

55


Sensor Network FeaturesSensor Network Features

APPLICATIONS: APPLICATIONS: Military, Environmental, Health, Home, Space Military, Environmental, Health, Home, Space Exploration, Exploration, Chemical Processing, Volcanoes, Mining, Disaster Chemical Processing, Volcanoes, Mining, Disaster ReliefRelief……. .

SENSOR TYPES:SENSOR TYPES:Seismic, Low Sampling Rate Magnetic, Thermal, Seismic, Low Sampling Rate Magnetic, Thermal, Visual, Infrared, Acoustic, Visual, Infrared, Acoustic, RadarRadar……

SENSOR TASKS:SENSOR TASKS:Temperature, Humidity, Vehicular Movement, Temperature, Humidity, Vehicular Movement, Lightning Condition,Lightning Condition,Pressure, Soil Makeup, Noise Levels, Presence or Pressure, Soil Makeup, Noise Levels, Presence or Absence of Certain Types of Objects, Mechanical Absence of Certain Types of Objects, Mechanical Stress Levels on Attached Objects, Current Stress Levels on Attached Objects, Current Characteristics Characteristics (Speed, Direction, Size) of an Object(Speed, Direction, Size) of an Object ……..


Examples for Sensor Examples for Sensor NodesNodes

Rockwell WINS

Smart Dust

JPL SensorWebs

Dust


MICA Mote

Rene Mote

weC Mote

Dot Mote

Examples for Sensor Examples for Sensor NodesNodes


Applications


Military ApplicationsMilitary Applications

Command, Control, Communications, Command, Control, Communications, ComputingComputing, Intelligence, Surveillance, , Intelligence, Surveillance, Reconnaissance, Reconnaissance, TargetingTargeting (C4ISRT)(C4ISRT)

Monitoring friendly forces, equipment and Monitoring friendly forces, equipment and ammunitionammunitionBattlefield surveillanceBattlefield surveillanceReconnaissance of opposing forces and terrainReconnaissance of opposing forces and terrainTargetingTargetingBattle damage assessmentBattle damage assessmentNuclear, Biological and Chemical (NBC) attack Nuclear, Biological and Chemical (NBC) attack detection and reconnaissancedetection and reconnaissance


Further Military Further Military ApplicationsApplications

Intrusion detection (mine fields)Intrusion detection (mine fields)Detection of firing gun (small arms) Detection of firing gun (small arms) locationlocationChemical (biological) attack detectionChemical (biological) attack detectionTargeting and target tracking systemsTargeting and target tracking systemsEnhanced navigation systemsEnhanced navigation systemsBattle damage assessment systemBattle damage assessment systemEnhanced logistics systemsEnhanced logistics systems


Environmental Environmental ApplicationsApplicationsTracking the movements of birds, small animals, and Tracking the movements of birds, small animals, and insectsinsectsMonitoring environmental conditions that affect crops Monitoring environmental conditions that affect crops and livestockand livestockIrrigationIrrigationEarth monitoring and planetary explorationEarth monitoring and planetary explorationChemical/biological detectionChemical/biological detectionBiological, Earth, and environmental monitoring in Biological, Earth, and environmental monitoring in marine, soil, and atmospheric contextsmarine, soil, and atmospheric contextsMeteorological or geophysical researchMeteorological or geophysical researchPollution studyPollution studyPrecision agriculturePrecision agricultureBiocomplexityBiocomplexity mapping of the environmentmapping of the environmentFlood detection, and Forest fire detection.Flood detection, and Forest fire detection.


Habitat MonitoringHabitat Monitoring

http://www.greatduckisland.netGreat Duck Island in Maine


Habitat MonitoringHabitat Monitoring

Approx. 200 nodes including MICA, MICA2, burrow Approx. 200 nodes including MICA, MICA2, burrow nodes (with IR) and weather station nodes nodes (with IR) and weather station nodes Motes detect light, barometric pressure, relative Motes detect light, barometric pressure, relative humidity and temperature conditions. humidity and temperature conditions. An infrared heat sensor detects whether the nest is An infrared heat sensor detects whether the nest is occupied by a seabird, and whether the bird has occupied by a seabird, and whether the bird has company. company. Motes within the burrows send readings out to a single Motes within the burrows send readings out to a single gateway sensor above ground, which then wirelessly gateway sensor above ground, which then wirelessly relays collected information to a laptop computer at a relays collected information to a laptop computer at a lighthouse (~350 feet). lighthouse (~350 feet). The laptop, also powered by photovoltaic cells, connects The laptop, also powered by photovoltaic cells, connects to the Internet via satellite.to the Internet via satellite.Computer at baseComputer at base--station logs data and maintains station logs data and maintains databasedatabase


Forest Fire Detection: Forest Fire Detection: FirebugFirebug

Design and Construction of a Wildfire Design and Construction of a Wildfire Instrumentation System using Instrumentation System using Networked SensorsNetworked SensorsNetwork of GPSNetwork of GPS--enabled, wireless enabled, wireless thermal sensors thermal sensors FireBugFireBug network selfnetwork self--organizes into organizes into edgeedge--hub configurations hub configurations Hub motes act are base stationsHub motes act are base stations

http://firebug.sourceforge.net/


FirebugFirebug

Firebug Firebug -- mote/fireboard mote/fireboard pair pair Mote Mote -- Crossbow MICA Crossbow MICA boardboardFireboard Fireboard -- Crossbow Crossbow MTS420CAMTS420CA

Temperature and Temperature and humidity sensor. humidity sensor. Barometric pressure Barometric pressure sensor. sensor. GPS unit. GPS unit. AccelerometerAccelerometerLight Intensity SensorLight Intensity Sensor


Observation and Forecasting Observation and Forecasting System for the Columbia River System for the Columbia River


Health ApplicationsHealth Applications

Providing interfaces for the disabledProviding interfaces for the disabledIntegrated patient monitoringIntegrated patient monitoringDiagnostics Diagnostics TelemonitoringTelemonitoring of human physiological of human physiological datadataTracking and monitoring doctors and Tracking and monitoring doctors and patients inside a hospitalpatients inside a hospitalDrug administration in hospitalsDrug administration in hospitals


CodeBlueCodeBlue: : WSNsWSNs for Medical for Medical Care Care

NSF, NIH, U.S. Army, Sun Microsystems and Microsoft NSF, NIH, U.S. Army, Sun Microsystems and Microsoft CorporationCorporationMotivation Motivation -- Vital sign data poorly integrated with preVital sign data poorly integrated with pre--hospital and hospitalhospital and hospital--based patient care recordsbased patient care records

http://www.eecs.harvard.edu/~mdw/proj/codeblue



HardwareHardwareSmall wearable sensorsSmall wearable sensorsWireless pulse Wireless pulse oximeteroximeter / 2/ 2--lead EKGlead EKGBased on the Mica2, Based on the Mica2, MicaZMicaZ, and , and TelosTelos sensor node sensor node platformsplatformsCustom sensor board with pulse Custom sensor board with pulse oximeteroximeter or EKG or EKG circuitrycircuitryPluto mote Pluto mote

scaledscaled--down version of the down version of the TelosTelosrechargeable Lirechargeable Li--ion batteryion batterysmall USB connectorsmall USB connector33--axis accelerometeraxis accelerometer



CodeBlueCodeBlue -- scalable software infrastructure for wireless scalable software infrastructure for wireless medical devicesmedical devices

Routing, Naming, Discovery, and Security Routing, Naming, Discovery, and Security MoteTrackMoteTrack -- tracking the location of individualtracking the location of individualpatient devices indoors and outdoorspatient devices indoors and outdoors

Heart rate (HR), oxygen saturation (SpO2), EKGHeart rate (HR), oxygen saturation (SpO2), EKGdata monitoreddata monitoredRelayed over a shortRelayed over a short--range (100m) range (100m) Receiving devices Receiving devices -- PDAsPDAs, laptops, ambulance, laptops, ambulance--based based terminalsterminalsData can be displayed in real time and integrated into the Data can be displayed in real time and integrated into the developing predeveloping pre--hospital patient care recordhospital patient care recordCan be programmed to process vital sign data (provide Can be programmed to process vital sign data (provide alerts)alerts)


Further ApplicationsFurther Applications

Monitoring product qualityMonitoring product qualityFactory Floor AutomationFactory Floor AutomationConstructing smart homesConstructing smart homesConstructing office spacesConstructing office spacesInteractive toysInteractive toysMonitor disaster areasMonitor disaster areasSmart spaces Smart spaces Machine diagnosisMachine diagnosisInteractive museumsInteractive museumsManaging inventory control Managing inventory control Environmental control in office buildingsEnvironmental control in office buildings


Disaster Relief Disaster Relief OperationsOperations

Drop sensor nodes from an aircraft over Drop sensor nodes from an aircraft over a WILDFIREa WILDFIRE

Each node measures temperatureEach node measures temperatureDerive a Derive a ““temperature maptemperature map””

Schools detect Schools detect airbornairborn toxins at low toxins at low concentrations, trace contaminant concentrations, trace contaminant transport to sourcetransport to sourceEarthquakeEarthquake--rubbledrubbled building infiltrated building infiltrated with robots and sensors: locate with robots and sensors: locate survivors, evaluate structural damagesurvivors, evaluate structural damage


Wireless Automatic Meter Reading Wireless Automatic Meter Reading SystemsSystems for Power Utilitiesfor Power Utilities


Wireless Automatic Meter Reading Wireless Automatic Meter Reading (WAMR) Systems(WAMR) Systems

Automatic meter reading functionalities: Automatic meter reading functionalities: RealReal--time energy consumption statisticstime energy consumption statisticsEffective billing managementEffective billing management

Telemetry functionalities: Telemetry functionalities: Remote control of equipmentRemote control of equipment

Dynamic configuration functionality: Dynamic configuration functionality: SelfSelf--configuration of the network in case of configuration of the network in case of route failures route failures

Status monitoring functionality: Status monitoring functionality: Monitoring the status of the metering devicesMonitoring the status of the metering devices


Buildings (or Bridges)Buildings (or Bridges)

HighHigh--rise buildings selfrise buildings self--detect detect structural faultsstructural faults

Reduce energy wastage by proper Reduce energy wastage by proper humidity, ventilation, air humidity, ventilation, air conditioning conditioning (HVAC) control (HVAC) control Needs measurements about room Needs measurements about room occupancy, temperature, air flow, occupancy, temperature, air flow, ……Monitor mechanical stress after Monitor mechanical stress after earthquakesearthquakes


More ApplicationsMore Applications

Facility ManagementFacility ManagementIntrusion detection into industrial sitesIntrusion detection into industrial sitesControl of leakages in chemical plants, Control of leakages in chemical plants, ……

Machine surveillance and preventive Machine surveillance and preventive maintenancemaintenance

Embed sensing/control functions into Embed sensing/control functions into places no cable has gone before places no cable has gone before E.g., tire pressure monitoringE.g., tire pressure monitoring


Underwater Sensor Underwater Sensor NetworksNetworks

Applications:Applications:Ocean Sampling Networks Ocean Sampling Networks Pollution Monitoring and other environmental Pollution Monitoring and other environmental monitoring (chemical, biological)monitoring (chemical, biological)Buoys alert swimmers to dangerous bacterial Buoys alert swimmers to dangerous bacterial levelslevelsDisaster Prevention Disaster Prevention Assisted NavigationAssisted NavigationDistributed Tactical SurveillanceDistributed Tactical SurveillanceMine ReconnaissanceMine Reconnaissance


Underwater Sensor NetworksUnderwater Sensor Networks3D Dynamic Architecture Using AUVs3D Dynamic Architecture Using AUVs

Drifters, Gliders


Communication architecture and solutions


Factors Influencing Factors Influencing Sensor Network DesignSensor Network Design

Fault Tolerance (Reliability)Fault Tolerance (Reliability)ScalabilityScalabilityProduction CostsProduction CostsHardware ConstraintsHardware ConstraintsSensor Network TopologySensor Network TopologyOperating EnvironmentOperating EnvironmentTransmission Media Transmission Media Power ConsumptionPower Consumption


Sensor Networks Sensor Networks Communication ArchitectureCommunication Architecture

Application LayerApplication Layer

Transport LayerTransport Layer

Network LayerNetwork Layer

Data Link LayerData Link Layer

Physical LayerPhysical Layer

Power M

anagement Plane

Mobility M

anagement Plane

Mobility M

anagement Plane

Task Managem

ent Plane

Used by sink and all sensor nodesUsed by sink and all sensor nodesCombines power and routing awarenessCombines power and routing awarenessIntegrates data with networking protocolsIntegrates data with networking protocolsCommunicates power efficiently throughCommunicates power efficiently through

wireless medium andwireless medium andPromotes cooperative efforts.Promotes cooperative efforts.


Why CanWhy Can’’t Adt Ad--hoc Network hoc Network Protocols Be Used Here?Protocols Be Used Here?

Number of sensor nodes can be several orders Number of sensor nodes can be several orders of magnitude higherof magnitude higherSensor nodes are densely deployed and are Sensor nodes are densely deployed and are prone to failuresprone to failuresThe topology of a sensor network may change The topology of a sensor network may change frequently due to node mobility and node frequently due to node mobility and node failurefailureSensor nodes are limited in power, Sensor nodes are limited in power, computational capacities, and memorycomputational capacities, and memoryMay not have global ID like IP addressMay not have global ID like IP addressNeed tight integration with sensing tasksNeed tight integration with sensing tasks


Application LayerApplication Layer


Sensor Network TopologySensor Network Topology

TaskTaskManagerManager

Users

Server(Database)

Internet, Internet, SatelliteSatellite


SMP: Sensor Management SMP: Sensor Management ProtocolProtocol

System Administrators interact with sensors via System Administrators interact with sensors via SMPSMPTasks:Tasks:

Turning sensors on and off Turning sensors on and off Moving the sensor nodesMoving the sensor nodesQuerying the sensor network configuration and the Querying the sensor network configuration and the status of nodes and restatus of nodes and re--configuring the sensor configuring the sensor networknetworkAuthentication, key distribution and securityAuthentication, key distribution and securityTimeTime--synchronization of sensor nodessynchronization of sensor nodesExchanging data related to location finding Exchanging data related to location finding algorithmsalgorithmsIntroducing rules related to data aggregation, Introducing rules related to data aggregation, attributeattribute--based naming, and clustering to sensor based naming, and clustering to sensor nodesnodes


Query ProcessingQuery Processing

Users can request data from the networkUsers can request data from the networkEfficient Query ProcessingEfficient Query Processing

User Query Types:User Query Types:Historical Queries:Historical Queries:

Used for analysis of historical data stored in a Used for analysis of historical data stored in a storage area storage area ““What was the temp. 2 hours back in the NW What was the temp. 2 hours back in the NW quadrant?quadrant?””

One Time Queries: One Time Queries: Gives a snapshot of the networkGives a snapshot of the network””What is the current temperature in the NW What is the current temperature in the NW quadrant?quadrant?””

Persistent Queries:Persistent Queries:Used to monitor the network over a time Used to monitor the network over a time interval with respect to some parametersinterval with respect to some parameters““Report the temperature for the next 2 hoursReport the temperature for the next 2 hours””


Sensor Query and Tasking Sensor Query and Tasking Language (SQTL)Language (SQTL)

•• SQTL is a procedural scripting languageSQTL is a procedural scripting language•• It provides interfaces to access sensor It provides interfaces to access sensor

hardware: hardware: getTemperaturegetTemperature, , turnOnturnOn

for location awareness:for location awareness:isNeighborisNeighbor, , getPositiongetPosition

and for communication:and for communication:tell execute.tell execute.

C-C Shen, et.al., “Sensor Information Networking Architecture and Applications”, IEEE Personal Communications Magazine, August 2001


SQTLSQTL

By using the upon command, a By using the upon command, a programmer can create an programmer can create an ““event event handling blockhandling block”” for 3 types of events:for 3 types of events:

Events generated when a message is Events generated when a message is received by a sensor node (RECEIVE)received by a sensor node (RECEIVE)Events triggered periodically (EVERY),Events triggered periodically (EVERY),Events caused by the expiration of a Events caused by the expiration of a timer (EXPIRE)timer (EXPIRE)


Task Assignment and Data Task Assignment and Data Advertisement ProtocolAdvertisement Protocol

Interest DisseminationInterest DisseminationUsers send their interest to a sensor Users send their interest to a sensor node, a subset of the nodes or the entire node, a subset of the nodes or the entire networknetworkThis interest may be about a certain This interest may be about a certain attribute of the sensor field or a attribute of the sensor field or a triggering eventtriggering event

Advertisement Of Available DataAdvertisement Of Available DataSensor nodes advertise the available Sensor nodes advertise the available data to the users and the users query data to the users and the users query the data which they are interested inthe data which they are interested in


Sensor Query and Data Sensor Query and Data Dissemination ProtocolDissemination Protocol

Provides user applications with Provides user applications with interfaces to issue queries, respond to interfaces to issue queries, respond to queries and collect incoming replies.queries and collect incoming replies.

Attribute Based QueryAttribute Based Query““The locations of the nodes that sense The locations of the nodes that sense temperature higher than 70Ftemperature higher than 70F””

Location Based QueryLocation Based Query““Temperatures read by the nodes in region Temperatures read by the nodes in region AA””


Attribute Based QueryAttribute Based Query

SinkSink

Query:Query:Sensor nodes that read >70Sensor nodes that read >70ooF temperatureF temperature

7171

6868

6868

6969

7171

7575

7171

6767

7171

6666


Location Based QueryLocation Based Query

SinkSink

Query:Query:Region A nodes should send their Region A nodes should send their temperaturestemperatures 7171

6868

6868

6969

7171

7575

7171

6767

7171

6666

Region ARegion A

Region BRegion B

Region CRegion C

Important for broadcasting, multicasting, geocasting and anycasting


Data Aggregation (Data Data Aggregation (Data Fusion)Fusion)

Query:Query:Sensor nodes that read >70Sensor nodes that read >70ooF temperatureF temperature

7171

7575

SinkSink

6868

6868

6969

71717171

6767

7171

6666

The sink asks the sensor nodes to report certain conditions.Data coming from multiple sensor nodes are aggregated.


Network layer


Energy Efficient RoutingEnergy Efficient Routing

SinkSink

E (PA=1)E (PA=1) F (PA=4)F (PA=4)

D (PA=3)D (PA=3)

A (PA=2)A (PA=2)B (PA=2)B (PA=2)

C (PA=2)C (PA=2)

TT

Route 1: SinkRoute 1: Sink--AA--BB--T (PA=4)T (PA=4)Route 2: SinkRoute 2: Sink--AA--BB--CC--T (PA=6)T (PA=6)Route 3: SinkRoute 3: Sink--DD--T (PA=3)T (PA=3)Route 4: SinkRoute 4: Sink--EE--FF--T (PA=5)T (PA=5)

Maximum power available (PA) routeMaximum power available (PA) routeMinimum hop routeMinimum hop routeMinimum energy routeMinimum energy routeMaximum minimum PA node Maximum minimum PA node routeroute (Route along which the (Route along which the minimum PA is larger than the minimum PA is larger than the minimum minimum PAsPAs of the other routesof the other routesis preferred, e.g., Route 3 is the is preferred, e.g., Route 3 is the most efficient; Route 1 is the most efficient; Route 1 is the second).second).


Routing in Routing in WSNsWSNs

Why not use conventional routing algorithms?Why not use conventional routing algorithms?Global (Unique) addresses, local addressesGlobal (Unique) addresses, local addressesUnique node addresses cannot be used in many Unique node addresses cannot be used in many sensor networkssensor networks

sheer number of nodes, energy constraints, data sheer number of nodes, energy constraints, data centric approachcentric approach

Node addressing is needed forNode addressing is needed fornode management, querying, data aggregation and node management, querying, data aggregation and fusion, service discovery, routingfusion, service discovery, routing

Very limited in power, computational capacities, Very limited in power, computational capacities, and memoryand memoryVery prone to failuresVery prone to failuresFrequent topology changeFrequent topology changeMay not have global ID like IP addressMay not have global ID like IP addressDensely deployedDensely deployedNeed tight integration with sensing tasksNeed tight integration with sensing tasks


Taxonomy of Routing Taxonomy of Routing ProtocolsProtocols

Data Centric ProtocolsData Centric ProtocolsFlooding, Gossiping, SPIN, SAR (Sequential Flooding, Gossiping, SPIN, SAR (Sequential Assignment Routing) , Directed Diffusion, Assignment Routing) , Directed Diffusion, Rumor Routing, ConstrainedRumor Routing, Constrained Anisotropic Diffused Anisotropic Diffused Routing, COUGAR, ACQUIRERouting, COUGAR, ACQUIRE

Hierarchical ProtocolsHierarchical ProtocolsLEACH, TEEN (Threshold Sensitive Energy LEACH, TEEN (Threshold Sensitive Energy Efficient Sensor Network Protocol), APTEEN, Efficient Sensor Network Protocol), APTEEN, PEGASISPEGASIS

Location Based ProtocolsLocation Based ProtocolsMECN, SMECN (Small Minimum Energy Com MECN, SMECN (Small Minimum Energy Com NetwNetw), GAF (Geographic Adaptive Fidelity), ), GAF (Geographic Adaptive Fidelity), GEAR, Distributed Topology/GEAR, Distributed Topology/GeographiGeographic c Routing Routing AlgorithmAlgorithm

K. K. AkkayaAkkaya and M. Younis, and M. Younis, ““A Survey on Routing Protocols for Wireless A Survey on Routing Protocols for Wireless Sensor Networks,Sensor Networks,”” AdHocAdHoc Networks (Elsevier) Networks (Elsevier) Journal, 2004Journal, 2004


DataData--Centric RoutingCentric Routing

Since sensor nodes are deployed Since sensor nodes are deployed randomly in largerandomly in large number, it is hard number, it is hard to assign specific IDs to each ofto assign specific IDs to each of the the sensor nodes. sensor nodes. Without a unique identifier, gathering Without a unique identifier, gathering data may become a challenge.data may become a challenge.To overcome this challenge, some To overcome this challenge, some routing protocolsrouting protocols gather/route data gather/route data based on the description of thebased on the description of the data, data, i.e., datai.e., data--centriccentric


DataData--Centric RoutingCentric Routing

The dataThe data--centric routing requires centric routing requires attribute based naming where the attribute based naming where the users are more interested in querying users are more interested in querying an attribute of the phenomenon, an attribute of the phenomenon, rather than querying an individual rather than querying an individual node. node. Example: Example:

"the areas where the temperature is over "the areas where the temperature is over 70F70F““ is a more common query than "the is a more common query than "the temperature read by a certain nodetemperature read by a certain node””


Data Data Dissemination Dissemination ApproachesApproaches

FloodingFloodingSend data to all neighborsSend data to all neighbors

GossipingGossipingSend data to one randomly Send data to one randomly selectedselectedneighborneighbor

Both are simple, butBoth are simple, but……ImplosionsImplosionsOverlapsOverlapsResource blindnessResource blindnessEnergy inefficiencyEnergy inefficiency

B

D E

FG

C

A


ProblemsProblems

A B

C (r,s)(q,r)

q s

Data OverlapData OverlapImplosionImplosion

A

B C

D

(a)

(a)

(a)

(a)

Resource BlindnessResource BlindnessNo knowledge about the available No knowledge about the available power of resourcespower of resources

r


GossipingGossiping

Uses randomization to save energyUses randomization to save energySelects a single node at random and Selects a single node at random and sends the data to itsends the data to itAvoids implosionsAvoids implosionsDistributes information slowlyDistributes information slowlyEnergy dissipates slowlyEnergy dissipates slowly


SPIN: Sensor Protocol for SPIN: Sensor Protocol for Information via NegotiationInformation via Negotiation

Two basic ideas:Two basic ideas:Sensors communicate with each other Sensors communicate with each other about the data that they already have about the data that they already have and the data they still need to obtainand the data they still need to obtain

to conserve energy and operate efficientlyto conserve energy and operate efficientlyexchanging data about sensor data may be exchanging data about sensor data may be cheapcheap

Sensors must monitor and adapt to Sensors must monitor and adapt to changes in their own energy resourceschanges in their own energy resources

W.R. Heinzelman, J. Kulik, and H. Balakrishnan, “Adaptive Protocols for Information Dissemination in Wireless Sensor Networks”, ACM MobiCom’99


SPINSPIN

Uses three types of messages: ADV, REQ, and Uses three types of messages: ADV, REQ, and DATADATAWhen a sensor node has something new, it When a sensor node has something new, it broadcasts an advertisement (ADV) packet that broadcasts an advertisement (ADV) packet that contains the new data, i.e., the meta datacontains the new data, i.e., the meta dataInterested nodes send a request (REQ) packet.Interested nodes send a request (REQ) packet.Data are sent to the nodes that request by Data are sent to the nodes that request by DATA packets.DATA packets.This will be repeated until all nodes will get a This will be repeated until all nodes will get a

copy.copy.


SPINSPIN

Good for disseminating information to all sensor nodes.Good for disseminating information to all sensor nodes.SPIN is based on dataSPIN is based on data--centric routing where the sensors broadcastcentric routing where the sensors broadcastan advertisement for the available data and wait for a requesan advertisement for the available data and wait for a request from t from interested sinksinterested sinks

1. ADV1. ADV

2. REQ2. REQ

3. DATA 3. DATA

Sensor Sensor


SPINSPIN

ADVADVREQREQ

DATADATA

ADVADVREQREQDATADATA


SPIN ProtocolSPIN Protocol

SPINSPIN33--stage handshake protocolstage handshake protocolAdvantagesAdvantages

Simple, Implosion avoidanceSimple, Implosion avoidance

Disadvantages Disadvantages Cannot isolate the nodes that do not want Cannot isolate the nodes that do not want to receive the information.to receive the information.Consume unnecessary powerConsume unnecessary power


Directed Diffusion Routing Directed Diffusion Routing AlgorithmAlgorithm

FeaturesFeaturesSink sends interest, i.e., task descriptor, to all sensor nodesSink sends interest, i.e., task descriptor, to all sensor nodes..Interest is named by assigning attributeInterest is named by assigning attribute--value pairs.value pairs.

sinksink

sourcesource

sinksink

sourcesource

sinksink

sourcesource

Interest PropagationInterest Propagation Gradient SetupGradient Setup Data DeliveryData Delivery



C. C. IntanagonwiwatIntanagonwiwat, , et.alet.al., ., ““Directed Diffusion: A Scalable and Robust Communication Directed Diffusion: A Scalable and Robust Communication Paradigm for Sensor NetworksParadigm for Sensor Networks””, IEEE/ACM Transactions on Networking, 2003., IEEE/ACM Transactions on Networking, 2003.

Interest PropagationInterest Propagation

SourceSource SinkSink

Gradient SetupGradient SetupData DeliveryData Delivery



This is a DATA CENTRIC ROUTING This is a DATA CENTRIC ROUTING schemeschemeMAIN IDEAMAIN IDEA

DIFFUSE data through sensor nodes DIFFUSE data through sensor nodes using a NAMING SCHEMEusing a NAMING SCHEME

REASONREASONGet rid off unnecessary operations of Get rid off unnecessary operations of routing schemes to SAVE ENERGY.routing schemes to SAVE ENERGY.


What is DATA CENTRIC?What is DATA CENTRIC?

DataData--CentricCentricSensor node does not need an identity!!!Sensor node does not need an identity!!!

What is the temp at node #27 ?What is the temp at node #27 ?Data is named by attributesData is named by attributes

Where are the nodes whose temperature Where are the nodes whose temperature recentlyrecently

exceeded 30 degrees ?exceeded 30 degrees ?How many pedestrians do you observe in How many pedestrians do you observe in

region X? region X? Tell me in what direction that vehicle in Tell me in what direction that vehicle in

region Y isregion Y ismoving?moving?

ApplicationApplication--SpecificSpecificNodes can perform application specific data Nodes can perform application specific data

aggregation, caching and forwardingaggregation, caching and forwarding


Directed DiffusionDirected Diffusion

NAMING SCHEME: Data generated by NAMING SCHEME: Data generated by sensor nodes is NAMED by ATTRIBUTEsensor nodes is NAMED by ATTRIBUTE--VALUE pairsVALUE pairsIn order to create a query, an interest is In order to create a query, an interest is defined using a list of attributedefined using a list of attribute--value value pairs such as name of objects, interval, pairs such as name of objects, interval, duration, geographical area, etc.duration, geographical area, etc.An arbitrary sensor node (usually the An arbitrary sensor node (usually the SINK) uses attributeSINK) uses attribute--value pairs value pairs (interests) for the data and queries the (interests) for the data and queries the sensors in an onsensors in an on--demand basisdemand basis


Example for AttributeExample for Attribute--Value Value PairsPairs

Example: (Animal Tracking Task)Example: (Animal Tracking Task)TypeType = four legged animal = four legged animal (detect animal location)(detect animal location)IntervalInterval = 20 ms = 20 ms (send back events every 20 ms)(send back events every 20 ms)DurationDuration = 10 seconds = 10 seconds (.. for the next 10 seconds)(.. for the next 10 seconds)RecRec = [= [--100,100,200,00] 100,100,200,00] (from sensors within the rectangle)(from sensors within the rectangle)

The task description specifies an interest for The task description specifies an interest for data matching for attributes data matching for attributes

called INTEREST.called INTEREST.


The data sent in response to interests are also named similarly.The data sent in response to interests are also named similarly.

Example: Example: Sensor detecting the animal generates the following data:Sensor detecting the animal generates the following data:

Type Type –– four legged animal four legged animal (type of animal seen)(type of animal seen)InstanceInstance= elephant = elephant (instance of this type)(instance of this type)LocatonLocaton = (125,220) = (125,220) (node location)(node location)IntensityIntensity = 0.6 = 0.6 (signal amplitude measure)(signal amplitude measure)ConfidenceConfidence = 085 = 085 (confidence in the match)(confidence in the match)TimestampTimestamp= 01:20:40 = 01:20:40 (event generation time)(event generation time)

Example for AttributeExample for Attribute--Value Value PairsPairs


Directed Diffusion Directed Diffusion OverviewOverviewThe sink broadcasts this interest to sensor nodes to The sink broadcasts this interest to sensor nodes to query information from a particular area in the query information from a particular area in the field.field.As the interest propagates, data may be locally As the interest propagates, data may be locally transformed (e.g., aggregated) at each node, or be transformed (e.g., aggregated) at each node, or be cachedcachedOther nodes express interests based on these Other nodes express interests based on these attributesattributesNetwork nodes propagate interestsNetwork nodes propagate interests

Interests establish gradients that direct diffusion of Interests establish gradients that direct diffusion of datadataPath of interest propagation sets up a reverse data Path of interest propagation sets up a reverse data path for data that matches the interestpath for data that matches the interest

The interests in the caches are then used to The interests in the caches are then used to compare the received data with the values in the compare the received data with the values in the interests. interests.


Rules for Interests and Rules for Interests and GradientsGradients

What are the local rules for propagating What are the local rules for propagating interests?interests?

E.g., just flood interestE.g., just flood interestMore sophisticated techniques possible: More sophisticated techniques possible: directional interest propagation, based on directional interest propagation, based on cached aggregate informationcached aggregate information

““I recently heard about suspicious activity from I recently heard about suspicious activity from neighbor A, so let me try sending this interest neighbor A, so let me try sending this interest for recent intrusions to that neighborfor recent intrusions to that neighbor””

What are the local rules for establishing What are the local rules for establishing gradients?gradients?

E.g., highest gradient towards neighbor who E.g., highest gradient towards neighbor who first sends interestfirst sends interestOthers possible e.g., towards neighbor with Others possible e.g., towards neighbor with highest remaining energyhighest remaining energy


Data Transmission Data Transmission ChoicesChoices

Different local data forwarding rules can Different local data forwarding rules can result in different kinds of transmissionresult in different kinds of transmission

single path deliverysingle path deliverymultimulti--path delivery, with traffic on each link path delivery, with traffic on each link proportional to its gradientproportional to its gradient

data probabilistically striped along different data probabilistically striped along different pathspathsredundant delivery across different pathsredundant delivery across different pathslayered transmission along different pathslayered transmission along different paths

delivery from single source to multiple sinksdelivery from single source to multiple sinksdelivery from multiple sources to multiple delivery from multiple sources to multiple sinkssinks


Directed Diffusion Directed Diffusion vsvsSPINSPIN

OnOn--Demand Data Query is differentDemand Data Query is differentIn DD In DD Sink queries sensors if a Sink queries sensors if a specific data is available by flooding specific data is available by flooding some taskssome tasksIn SPIN In SPIN Sensors advertise the Sensors advertise the availability of data allowing sinks to availability of data allowing sinks to query that data.query that data.


Directed Diffusion Directed Diffusion AdvantagesAdvantages

DD is data centric DD is data centric no need for a no need for a node addressing mechanismnode addressing mechanismEach node can do aggregation, Each node can do aggregation, caching and sensingcaching and sensingDD is energy efficient since it is on DD is energy efficient since it is on demand and no need to maintain demand and no need to maintain global network topology.global network topology.


Directed Diffusion Directed Diffusion DisadvantagesDisadvantages

Not generally applicable since it is based Not generally applicable since it is based on a query driven data delivery model. on a query driven data delivery model. For applications needing continuous data For applications needing continuous data delivery (e.g., environmental delivery (e.g., environmental monitoring) monitoring) DD is not a good choice.DD is not a good choice.Naming schemes are application Naming schemes are application dependent and each time must be dependent and each time must be defined adefined a--priori.priori.Matching process for data and queries Matching process for data and queries cause some overhead at sensors.cause some overhead at sensors.


Hierarchical ProtocolsHierarchical Protocols

Sensor nodes are deployed with a limited amount of Sensor nodes are deployed with a limited amount of energyenergyHierarchicalHierarchical--architecture protocols are proposed to architecture protocols are proposed to address the scalability and energy consumption address the scalability and energy consumption challenges of sensor networkschallenges of sensor networksSensor nodes form clusters where the clusterSensor nodes form clusters where the cluster--heads heads aggregate and fuse data to conserve energyaggregate and fuse data to conserve energyThe clusterThe cluster--heads may form another layer of clusters heads may form another layer of clusters among themselves before reaching the sink. among themselves before reaching the sink. ExamplesExamples

LowLow--Energy Adaptive Clustering Hierarchy (LEACH) Energy Adaptive Clustering Hierarchy (LEACH) (Heinzelman(Heinzelman’’00)00)PowerPower--efficient efficient GAtheringGAthering in Sensor Information in Sensor Information Systems (PEGASIS) (Lindsey01, Lindsey02), Systems (PEGASIS) (Lindsey01, Lindsey02), Threshold sensitive Energy Efficient sensor Network Threshold sensitive Energy Efficient sensor Network protocol (TEEN) (Manjeshwar01)protocol (TEEN) (Manjeshwar01)AdaPtiveAdaPtive Threshold sensitive Energy Efficient sensor Threshold sensitive Energy Efficient sensor Network Protocol (APTEEN) (Manjeshwar02)Network Protocol (APTEEN) (Manjeshwar02)


Low Energy Adaptive Clustering Low Energy Adaptive Clustering Hierarchy (LEACH)Hierarchy (LEACH)

LEACH is a clustering based protocol LEACH is a clustering based protocol which minimizes energy dissipation in which minimizes energy dissipation in sensor networkssensor networksIdea: Idea:

Randomly select sensor nodes as cluster Randomly select sensor nodes as cluster heads, so the high energy dissipation in heads, so the high energy dissipation in communicating with the base station is spread communicating with the base station is spread to all sensor nodes in the networkto all sensor nodes in the networkForming clusters is based on the received Forming clusters is based on the received signal strengthsignal strengthCluster heads can then be used kind of routers Cluster heads can then be used kind of routers (relays) to the sink.(relays) to the sink.

W. R. Heinzelman, A. Chandrakasan, and H. Balakrishnan, “Energy-Efficient Communication Protocol for Wireless Microsensor Networks,'' IEEE Proc. of the Hawaii Int. Conf. on System Sciences, January 2000.


LEACHLEACH

Two Phases: SetTwo Phases: Set--up Phase and Steadyup Phase and Steady--PhasePhaseSet-up Phase:* Sensors may elect themselves to be a local cluster head at any time with a certain probability. (Reason: to balance the energy dissipation)* A sensor node chooses a random number between 0 and 1.* If this random number is less than the threshold T(n), the sensor node becomes a cluster-head.

T(n) = P / {1 – P[r mod (1/P)]} if n is element of Gwhere P is the desired percentage to become a cluster head (e.g., 0.05)

r is the current roundG is the set of nodes that have not been a cluster head in the

last 1/P rounds.


LEACHLEACH

* After the cluster heads are selected, the cluster * After the cluster heads are selected, the cluster heads advertise to all sensor nodes in the network heads advertise to all sensor nodes in the network that they are the new cluster heads.that they are the new cluster heads.

* Each node accesses the network through the cluster * Each node accesses the network through the cluster head that requires minimum energy to reach.head that requires minimum energy to reach.


Dynamic ClustersDynamic Clusters


LEACHLEACH

Once the nodes receive the Once the nodes receive the advertisement, they determine the advertisement, they determine the cluster that they want to belong based cluster that they want to belong based on the signal strength of the on the signal strength of the advertisement from the cluster heads to advertisement from the cluster heads to the sensor nodes.the sensor nodes.The nodes inform the appropriate cluster The nodes inform the appropriate cluster heads that they will be a member of the heads that they will be a member of the clusterclusterAfterwards the cluster heads assign the Afterwards the cluster heads assign the time on which the sensor nodes can time on which the sensor nodes can send data to themsend data to them


LEACHLEACH

STEADY STATE PHASE:STEADY STATE PHASE:

Sensors begin to sense and transmit data to the cluster headsSensors begin to sense and transmit data to the cluster headswhich aggregate data from the nodes in their clusters.which aggregate data from the nodes in their clusters.

After a certain period of time spent on the steady state,After a certain period of time spent on the steady state,the network goes into startthe network goes into start--up phase again and enters another up phase again and enters another round of selecting cluster heads.round of selecting cluster heads.


LEACHLEACH

Optimum Number of Clusters Optimum Number of Clusters ------????????????????

-- too few: nodes far from cluster too few: nodes far from cluster headsheads

too many: many nodes send data to too many: many nodes send data to SINK.SINK.


LEACHLEACH

Achieves over a factor of 7 reduction in energy Achieves over a factor of 7 reduction in energy dissipation compared to direct communication.dissipation compared to direct communication.The nodes die randomly and dynamic clustering The nodes die randomly and dynamic clustering increases lifetime of the system.increases lifetime of the system.It is completely distributed and requires no global It is completely distributed and requires no global knowledge of the network.knowledge of the network.It uses single hop routing where each node can It uses single hop routing where each node can transmit directly to the cluster head and the sink.transmit directly to the cluster head and the sink.It is not applicable to networks deployed in large It is not applicable to networks deployed in large regions.regions.Furthermore, the idea of dynamic clustering brings Furthermore, the idea of dynamic clustering brings extra overhead, e.g., head changes, extra overhead, e.g., head changes, advertisements etc. which may diminish the gain in advertisements etc. which may diminish the gain in energy consumption.energy consumption.


LocationLocation--Based ProtocolsBased Protocols

If the locations of the sensor nodes If the locations of the sensor nodes are known, the routing protocols can are known, the routing protocols can use this information to reduce the use this information to reduce the latency and latency and energy consumption of the sensor energy consumption of the sensor network.network.Although GPS is not envisioned for all Although GPS is not envisioned for all types of sensor networks, it can still types of sensor networks, it can still be used if stationary nodes with large be used if stationary nodes with large amount of energy are allowedamount of energy are allowed


Example: Minimum Energy Example: Minimum Energy Communication NetworkCommunication Network

8383

Minimum Energy Communication Network (MECN), L. Li and J.Y. Minimum Energy Communication Network (MECN), L. Li and J.Y. HalpernHalpern, , ““MinimumMinimum--Energy Mobile Wireless Networks RevisitedEnergy Mobile Wireless Networks Revisited”” IEEE ICC 2001IEEE ICC 2001

Uses graph theory:Uses graph theory:* Each node knows its exact location* Each node knows its exact location* Network is represented by a graph G* Network is represented by a graph G’’, and it is assumed that the, and it is assumed that the

resulting graph is connected resulting graph is connected * A sub* A sub--graph G of Ggraph G of G’’ is computed. is computed. * G connects all nodes with minimum energy cost.* G connects all nodes with minimum energy cost.

AA

BB

Connection A requires less energy than Connection A requires less energy than connection B because the power required to connection B because the power required to transmit between a pair of nodes increases as transmit between a pair of nodes increases as the nthe nthth power of the distance between them power of the distance between them (n>=2).(n>=2).


MAC


Objectives of MAC Objectives of MAC ProtocolsProtocols

Collision AvoidanceCollision AvoidanceEnergy EfficiencyEnergy EfficiencyScalabilityScalabilityLatencyLatencyFairnessFairnessThroughputThroughputBandwidth UtilizationBandwidth Utilization


Power ConsumptionPower Consumption

SENSOR

0

5

10

15

20

Pow

er (m

W)

CPU TX RX IDLE SLEEP

RADIO


Major Sources of Energy Major Sources of Energy WasteWaste

* Idle Listening- Long idle time when no sensing event happens- Collisions- Control Overhead- Overhearing

* Transmitter* Receiver

OBJECTIVE: Reduce energy consumption !!

Common to all wireless networks


Challenges in Medium Challenges in Medium Access Control in Access Control in WSNsWSNs

1. WSN Architecture1. WSN ArchitectureHigh density of nodesHigh density of nodesIncreased collision probabilityIncreased collision probabilitySignaling overhead should be minimized Signaling overhead should be minimized to prevent further collisionsto prevent further collisionsSophisticated and simple collision Sophisticated and simple collision avoidance protocols requiredavoidance protocols required


Challenges in Medium Access Challenges in Medium Access Control in Control in WSNsWSNs

2. Limited Energy Resources2. Limited Energy ResourcesConnectivity and the performance of the Connectivity and the performance of the network is affected as nodes dienetwork is affected as nodes dieTransmitting and receiving consumes almost Transmitting and receiving consumes almost same energysame energyFrequent power up/down eats up energyFrequent power up/down eats up energyNeed very low power MAC protocolsNeed very low power MAC protocolsMinimize signaling overheadMinimize signaling overheadAvoid idle listeningAvoid idle listeningPrevent frequent radio state changes Prevent frequent radio state changes ((activeactive sleepsleep))



3. Limited Processing and Memory 3. Limited Processing and Memory CapabilitiesCapabilities

Complex algorithms cannot be implementedComplex algorithms cannot be implementedConventional layered architecture may not bConventional layered architecture may not be e appropriateappropriateCentralized or local management is limitedCentralized or local management is limitedSimple scheduling algorithms requiredSimple scheduling algorithms requiredCrossCross--layer optimization requiredlayer optimization requiredSelfSelf--configurable, distributed protocols configurable, distributed protocols requiredrequired



4. Limited Packet Size4. Limited Packet SizeUnique node ID is not practicalUnique node ID is not practicalLimited header spaceLimited header spaceLocal IDs should be used for interLocal IDs should be used for inter--node node communicationcommunicationMAC protocol overhead should be minimizedMAC protocol overhead should be minimized

5. Cheap Encoder/Decoders5. Cheap Encoder/DecodersCheap node requirement prevents Cheap node requirement prevents

sophisticatedsophisticated encoders/decoders to be encoders/decoders to be implementedimplementedSimple FEC codes required for error controlSimple FEC codes required for error controlChannel state dependent MAC can be used toChannel state dependent MAC can be used to

decreasedecrease error rateerror rate



6. Inaccurate Clock Crystals6. Inaccurate Clock CrystalsCheap node requirement prevents expensive Cheap node requirement prevents expensive crystalscrystals to be implementedto be implementedSynchronization problemsSynchronization problemsTDMATDMA--based schemes are not practicalbased schemes are not practical

7. Event7. Event--based Networkingbased NetworkingObserved data depends on physical Observed data depends on physical phenomenonphenomenonSpatial and temporal correlation in the Spatial and temporal correlation in the physicalphysical phenomenon should be exploitedphenomenon should be exploited

BOTTOMLINE: Existing MAC protocols cannot be used for BOTTOMLINE: Existing MAC protocols cannot be used for sensor networks!!!!sensor networks!!!!


Overview of MAC Overview of MAC ProtocolsProtocols

1. Contention1. Contention--Based MAC ProtocolsBased MAC Protocols

2. Reservation2. Reservation--Based MAC ProtocolsBased MAC Protocols

3. HYBRID MAC PROTOCOLS3. HYBRID MAC PROTOCOLS

4. CROSS LAYER PROTOCOLS4. CROSS LAYER PROTOCOLS


ContentionContention--Based MAC Based MAC ProtocolsProtocols

Every node tries to access the channel Every node tries to access the channel based on carrier sense mechanismbased on carrier sense mechanismContentionContention--based protocols provide based protocols provide robustness and scalability to the robustness and scalability to the networknetworkThe collision probability increases with The collision probability increases with increasing increasing nnode densityode densityThey can support variable, but highly They can support variable, but highly correlated and dominantly periodic correlated and dominantly periodic traffic.traffic.


SMACSMAC

Aims to decrease energy consumption Aims to decrease energy consumption by sleep schedulesby sleep schedulesVirtual clustering of sensor nodesVirtual clustering of sensor nodesOverhear avoidance and message Overhear avoidance and message passingpassingTrades off latency for reduced energy Trades off latency for reduced energy consumptionconsumptionNeighbor discovery and channel Neighbor discovery and channel assignment combinedassignment combinedRandom wake up of sensors from sleep Random wake up of sensors from sleep schedulesschedules

W. Ye, et. al., “Medium Access Control with Coordinated Adaptive Sleeping for Wireless Sensor Networks,‘‘ IEEE/ACM Trans. on Networking, June 2004.


SS--MACMAC

Trade Off Trade Off Energy Efficiency versus NodeEnergy Efficiency versus Node--Level Fairness & Level Fairness & LatencyLatency

Basic Scheme: Periodic Listen and SleepBasic Scheme: Periodic Listen and SleepChoosing ScheduleChoosing Schedule

The node randomly chooses a time to go to sleep.The node randomly chooses a time to go to sleep.The node receives and follows its neighborThe node receives and follows its neighbor’’s s schedule by setting its schedule to be the same.schedule by setting its schedule to be the same.If the node receives a different schedule after it If the node receives a different schedule after it selects its own schedule, it adopts both schedulesselects its own schedule, it adopts both schedules

Maintaining ScheduleMaintaining ScheduleTo update schedule by sending a SYNC packet To update schedule by sending a SYNC packet periodically periodically


SS--MACMAC

Collision AvoidanceCollision AvoidanceVirtual carrier senseVirtual carrier sense

Checking the keeping silent time recorded Checking the keeping silent time recorded in the NAV (network allocation vector) in the NAV (network allocation vector)

Physical carrier sensePhysical carrier senseListening to the channel for possible Listening to the channel for possible transmissiontransmission

RTS/CTS exchangeRTS/CTS exchange

Overhearing AvoidanceOverhearing AvoidanceLet interfering nodes go to sleep after Let interfering nodes go to sleep after they hear an RTS or CTS packet they hear an RTS or CTS packet


SS--MACMAC

Message passingMessage passingOnly one RTS packet and one CTS packet are Only one RTS packet and one CTS packet are usedused

To avoid large control overhead and long To avoid large control overhead and long delaydelay

ACK would be sent after each data fragmentACK would be sent after each data fragmentTo avoid fragment loss or errorTo avoid fragment loss or errorTo prevent hidden terminal problem To prevent hidden terminal problem

After the neighbor node hears the RTS and After the neighbor node hears the RTS and CTS, it will go to sleep for the time that is CTS, it will go to sleep for the time that is needed to transmit all the fragments (using needed to transmit all the fragments (using the duration field) the duration field)


SS--MACMAC

Energy Saving vs. Increasing Latency (MultiEnergy Saving vs. Increasing Latency (Multi--hop Network)hop Network)Carrier sense delayCarrier sense delay

Determined by the contention window size Determined by the contention window size BackoffBackoff delaydelay

Because the node detects another transmission or the collision Because the node detects another transmission or the collision occurs occurs

Transmission delayTransmission delayDetermined by channel bandwidth, packet length and the coding Determined by channel bandwidth, packet length and the coding scheme adoptedscheme adopted

Propagation delayPropagation delayDetermined by the distance between the sending and receiving Determined by the distance between the sending and receiving nodesnodes

Processing delayProcessing delayDepends on the computing power of the node and the efficiency Depends on the computing power of the node and the efficiency of inof in--network data processing algorithms network data processing algorithms

QueueingQueueing delaydelayDepends on the traffic loadDepends on the traffic load

Sleep delaySleep delayCaused by the nodes periodic sleeping Caused by the nodes periodic sleeping


SS--MACMAC

SMAC has very good energy conserving SMAC has very good energy conserving compared with IEEE802.11compared with IEEE802.11ShortcomingsShortcomings

Nodes belonging to different subnets might not Nodes belonging to different subnets might not be able to connect be able to connect A mainly static network is assumedA mainly static network is assumed

Trades off latency for reduced energy Trades off latency for reduced energy consumptionconsumptionRedundant data is still sent with increased Redundant data is still sent with increased latencylatencyIncreased collision rate due to sleep schedules Increased collision rate due to sleep schedules


TT--MACMAC

One of the disadvantages of SOne of the disadvantages of S--MAC is that it can MAC is that it can not provide not provide adaptivityadaptivity to to burstybursty traffic since the traffic since the sleep schedules are fixed lengthsleep schedules are fixed lengthTT--MAC introduces an adaptive duty cycle. MAC introduces an adaptive duty cycle. The nodes listen to the channel only when there is The nodes listen to the channel only when there is traffic which reduces the amount of energy wasted traffic which reduces the amount of energy wasted on idle listeningon idle listeningVariable sleep schedules are introduced for Variable sleep schedules are introduced for different trafficdifferent trafficBoth SBoth S--MAC and TMAC and T--MAC provide significant energy MAC provide significant energy savings compared to IEEE 802.11, however, at the savings compared to IEEE 802.11, however, at the cost of increased latency and throughput cost of increased latency and throughput degradationdegradation

T. van Dam, K. Langendoen, ``An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks,’’ Proc. ACM SenSys, November 2003


Remarks On Contention Remarks On Contention Based ProtocolsBased Protocols

Generally, contentionGenerally, contention--based protocols provide based protocols provide scalability and lower delay, when compared to scalability and lower delay, when compared to reservationreservation--based protocolsbased protocolsOn the other hand, the energy consumption is On the other hand, the energy consumption is significantly higher than the TDMAsignificantly higher than the TDMA--based approaches based approaches due to collisions and collision avoidance schemesdue to collisions and collision avoidance schemesMoreover, contentionMoreover, contention--based protocols are more based protocols are more adaptive to the changes in the traffic volume and adaptive to the changes in the traffic volume and hence applicable to applications with hence applicable to applications with burstybursty traffic traffic such as eventsuch as event--based applicationsbased applicationsFurthermore, the synchronization and clustering Furthermore, the synchronization and clustering requirements of reservationrequirements of reservation--based protocols make based protocols make contentioncontention--based more favorable in scenarios where based more favorable in scenarios where such requirements can not be fulfilled.such requirements can not be fulfilled.


Reservation Based MAC Reservation Based MAC ProtocolsProtocolsThese are collisionThese are collision--free MAC protocols!free MAC protocols!The advantage of collisionThe advantage of collision--free communication free communication since each node transmits data to a central since each node transmits data to a central agent during its reserved slotagent during its reserved slotHence, the duty cycle of the nodes is Hence, the duty cycle of the nodes is decreased resulting in further energy efficiencydecreased resulting in further energy efficiencyGenerally, these protocols follow common Generally, these protocols follow common principles where the network is divided into principles where the network is divided into clusters and each node communicates clusters and each node communicates according to a specific superaccording to a specific super--frame structureframe structure


Reservation Based MAC Reservation Based MAC ProtocolsProtocols

The superThe super--frame structure consists of frame structure consists of two main partstwo main parts

1. Reservation Period is used by the nodes to 1. Reservation Period is used by the nodes to reserve their slots for communication reserve their slots for communication through a central agent, i.e., clusterthrough a central agent, i.e., cluster--headhead2. Data Period consists of multiple slots, that 2. Data Period consists of multiple slots, that is used by each sensor for transmitting is used by each sensor for transmitting informationinformation

Among the proposed TDMA schemes, Among the proposed TDMA schemes, the contention schemes for reservation the contention schemes for reservation protocols, the slot allocation principles, protocols, the slot allocation principles, the frame size and clustering the frame size and clustering approaches differ in each protocolapproaches differ in each protocol


TRAMA: Energy Efficient TRAMA: Energy Efficient CollisionCollision--Free MACFree MAC

The protocol is based on a timeThe protocol is based on a time--slotted slotted structure and uses a distributed election structure and uses a distributed election scheme based on traffic requirements of each scheme based on traffic requirements of each node to determine the time slot that a node node to determine the time slot that a node should useshould useEach node gets information about its every Each node gets information about its every twotwo--hop neighbor and the traffic information of hop neighbor and the traffic information of each node during a random access period, i.e., each node during a random access period, i.e., the reservation periodthe reservation periodBased on this information, each node Based on this information, each node calculates its priority and decides on which calculates its priority and decides on which time slot to usetime slot to useNodes sleep during their allocated slots if they Nodes sleep during their allocated slots if they do not have any packets to send or receivedo not have any packets to send or receive

V. Rajendran, K. Obraczka, and J. J. Garcia-Luna-Aceves,``Energy-Efficient, Collision-Free Medium Access Control for Wireless Sensor Networks,'' Proc. ACM SenSys 2003, LA, CA, Nov. 2003


BitBit--MapMap--Assisted (BMA) MAC Assisted (BMA) MAC ProtocolProtocol

An intraAn intra--cluster communication bitcluster communication bit--mapmap--assisted assisted (BMA) MAC protocol with an energy efficient TDMA (E(BMA) MAC protocol with an energy efficient TDMA (E--TDMA) schemeTDMA) schemeThe protocol consists of cluster setThe protocol consists of cluster set--up phase and up phase and steady state phasesteady state phaseIn the cluster setIn the cluster set--up phase, cluster head is selected up phase, cluster head is selected based on the available energy in each nodebased on the available energy in each nodeAccordingly, an EAccordingly, an E--TDMA MAC scheme is used in each TDMA MAC scheme is used in each of the clusters formed by the clusterof the clusters formed by the cluster--headsheadsIn each super frame, the reservation period is slotted In each super frame, the reservation period is slotted for contention and the data period is divided into two for contention and the data period is divided into two periods, i.e., data transmission period and idle periodperiods, i.e., data transmission period and idle periodThe duration of the data period is fixed and the data The duration of the data period is fixed and the data transmission period is changed based on the traffic transmission period is changed based on the traffic demands of the nodesdemands of the nodes

J. Li, G. Y. Lazarou, ``A bit-map-assisted energy-efficient MAC Scheme for wireless sensor networks,'' Proc. ACM IPSN, April 2004


Remarks OnRemarks OnReservation Based ProtocolsReservation Based Protocols

Overall, Contention Based Protocols, i.e., TDMAOverall, Contention Based Protocols, i.e., TDMA--based protocols provide collisionbased protocols provide collision--free free communication and achieve improved energy communication and achieve improved energy efficiencyefficiencyHowever, such TDMAHowever, such TDMA--based protocols require an based protocols require an infrastructure consisting of cluster heads which infrastructure consisting of cluster heads which coordinate the time slots assigned to each nodecoordinate the time slots assigned to each nodeAlthough many clustering algorithms have been Although many clustering algorithms have been proposed with these protocols, the optimality and proposed with these protocols, the optimality and the energy efficiency of these algorithms still need the energy efficiency of these algorithms still need to be investigatedto be investigatedIn addition, TDMAIn addition, TDMA--based protocols cause high based protocols cause high latency due to the frame structurelatency due to the frame structure


Remarks OnRemarks On Reservation Reservation Based ProtocolsBased Protocols

Hence TDMAHence TDMA--based MAC protocols may not be based MAC protocols may not be suitable for WSN applications where delay is suitable for WSN applications where delay is important in estimating event features and the important in estimating event features and the traffic has traffic has burstybursty naturenatureMoreover, since a time slotted communication Moreover, since a time slotted communication is performed in the clusters, interis performed in the clusters, inter--cluster cluster interference has to be minimized such that interference has to be minimized such that nodes with overlappingnodes with overlapping schedules in different schedules in different clusters do not collide which each otherclusters do not collide which each otherFinally, time synchronization is an important Finally, time synchronization is an important part of the TDMApart of the TDMA--based protocols and based protocols and synchronization algorithms are requiredsynchronization algorithms are required


Hybrid Medium AccessHybrid Medium AccessControl SchemesControl Schemes

While a pure TDMAWhile a pure TDMA--based access scheme based access scheme dedicates the entire channel to a single sensor dedicates the entire channel to a single sensor node, a pure Frequencynode, a pure Frequency--Division Multiple Division Multiple Access (FDMA) or CodeAccess (FDMA) or Code--Division Multiple Division Multiple Access (CDMA) scheme allocates minimum Access (CDMA) scheme allocates minimum signal bandwidth per nodesignal bandwidth per nodeSuch contrast brings the tradeoff between the Such contrast brings the tradeoff between the access capacity and the energy consumptionaccess capacity and the energy consumptionHybrid schemes in reservationHybrid schemes in reservation--based protocols based protocols aim to leverage the tradeoff introduced in aim to leverage the tradeoff introduced in channel allocation by combining TDMA channel allocation by combining TDMA approaches with FDMA or CDMA schemesapproaches with FDMA or CDMA schemes


Hybrid TDMA/FDMAHybrid TDMA/FDMA

An analytical formula is derived to find An analytical formula is derived to find the optimum number of channels for the optimum number of channels for each node, which gives the minimum each node, which gives the minimum system power consumption. system power consumption. This determines the hybrid TDMAThis determines the hybrid TDMA--FDMA FDMA scheme to be usedscheme to be usedAlthough the MAC protocol also assumes Although the MAC protocol also assumes that clusters are formed in the network, that clusters are formed in the network, it is a good example of crossit is a good example of cross--layer layer optimization where the MAC protocol is optimization where the MAC protocol is designed according to the physical layer designed according to the physical layer properties.properties.E.Shih, et al., Physical Layer Driven Protocol and Algorithm Design for Energy-Efficient Wireless Sensor Networks, Proc. of ACM Mobicom'01, July 2001


Remarks OnRemarks On Hybrid Hybrid ProtocolsProtocolsADVANTAGES:ADVANTAGES:

The hybrid reservation protocols provide The hybrid reservation protocols provide performance enhancements in terms of performance enhancements in terms of collision avoidance and energycollision avoidance and energy efficiency due efficiency due to improved channel organizationto improved channel organization

DISADVANTAGES:DISADVANTAGES:They require sophisticated physical and MAC They require sophisticated physical and MAC layer protocols that support CDMA or FDMA layer protocols that support CDMA or FDMA communication or unique radio componentscommunication or unique radio componentsHence, these protocols may not be applicable Hence, these protocols may not be applicable for high density WSN where sensor node cost for high density WSN where sensor node cost is an important factoris an important factor


Cross Layer SolutionsCross Layer Solutions

Incorporating physical layer and network layer Incorporating physical layer and network layer information into the MAC layer design improves the information into the MAC layer design improves the performance in WSNperformance in WSNRouteRoute--aware protocols provide lower delay bounds aware protocols provide lower delay bounds while physical layer coordination improve the while physical layer coordination improve the energy efficiency of the overall systemenergy efficiency of the overall systemMoreover, since sensor nodes are characterized by Moreover, since sensor nodes are characterized by their limitedtheir limited energy capabilities and memory energy capabilities and memory capacities, crosscapacities, cross--layer solutionslayer solutions provide efficient provide efficient solutions in terms of both performance and costsolutions in terms of both performance and costHowever, care must be taken while designing However, care must be taken while designing crosscross--layer solutions since the interdependence of layer solutions since the interdependence of each parameter should be analyzed in detaileach parameter should be analyzed in detail


CrossCross--Layer SolutionsLayer Solutions

MultiMulti--hop introduces a network delay in Shop introduces a network delay in S--MAC. MAC. Latency increases linearly with the hop count. Latency increases linearly with the hop count. Although the slope of the linearity is reduced to half Although the slope of the linearity is reduced to half by the adaptive listening algorithm, the protocol by the adaptive listening algorithm, the protocol still introduces significant latency compared to pure still introduces significant latency compared to pure contentioncontention--based protocols. based protocols. Also the frame structure in TDMAAlso the frame structure in TDMA--based protocols based protocols plus multiplus multi--hopping imposes significant network hopping imposes significant network delay. delay. MultiMulti--hop route of the packets should be considered hop route of the packets should be considered in MACin MACConsequently, routeConsequently, route--aware protocols have been aware protocols have been proposedproposed

W. C. Park et.al., ``Trade-off Energy and Delay between MAC Protocols for Wireless Sensor Networks,‘‘ in Proc. Int. Conf. on Advanced Communication Technology, Feb’04


DMAC: Data Gathering DMAC: Data Gathering MACMACA routeA route--aware contentionaware contention--based MAC protocol for data based MAC protocol for data gatheringgatheringData is collected through a unidirectional treeData is collected through a unidirectional treeDMAC introduces a sleep schedule such that the DMAC introduces a sleep schedule such that the nodes on a nodes on a multihopmultihop path wake up sequentially as the path wake up sequentially as the packet traversespacket traversesMoreover, since small sized packets are used, Moreover, since small sized packets are used, RTS/CTS mechanism is not used. RTS/CTS mechanism is not used. DMAC incorporates local synchronization protocols in DMAC incorporates local synchronization protocols in order to perform local scheduling and uses data order to perform local scheduling and uses data prediction in case a node requires a higher duty cycle prediction in case a node requires a higher duty cycle for data transmissionfor data transmissionBased on these techniques, multiBased on these techniques, multi--hop effects on the hop effects on the delay performance is minimized specifically for data delay performance is minimized specifically for data gathering applications where a unidirectional tree is gathering applications where a unidirectional tree is usedused“An Adaptive Energy-Efficient and Low-Latency MAC for Data Gathering in Wireless Sensor Networks”, G. Lu, B. Krishnamachari, C.S. Raghavendra, Proc. IEEE Int. Parallel and Distributed Processing Symposium, April 2004

9 – Sensor Networks - Politecnico di Milanohome.deib.polimi.it/capone/rmd/9-Reti di...

Documents

Transcript of 9 – Sensor Networks - Politecnico di Milanohome.deib.polimi.it/capone/rmd/9-Reti di...