Post on 14-Jun-2015
description
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Cloud Computing and Intelligent Systems
Two Fields at a Crossroads
Dr. Jeffrey Wallace
The 2014 International Conference on Computational Science and Computational Intelligence (CSCI'14)
Infinite Dimensions
Solving Tomorrow’s Problems Today
2
CSCI'14
Cloud Computing and Intelligent Systems
• Context• Technical Challenges• Examples
– Unmanned Systems Control• Anti-ship Missile Defense• Battlefield Extraction of Wounded
• Technology Enablers– System/Component Integration– Algorithm/Environment Integration
• Summary
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Context• Unique Systems Engineering and Development
Capability– A distillation of over $2B in government R&D– Addresses both people and technology
• Partnered with the Office of the Secretary of Defense, the Joint Chiefs of Staff, the 4 services, Academia, and others
• Solved world-class/grand challenge problems on F-35, CVN-21, healthcare IT, etc.
• Systems Integrator for the Continuous Transformation Environment
3
Infinite Dimensions
Solving Tomorrow’s Problems Today
Continuous Transformation Environment
“A Cloud Computing Range”
Infinite Dimensions
Solving Tomorrow’s Problems Today
5
CSCI'14
CTE at a Glance• CTE Objectives
– Prototype development and experimentation: Innovation in a collaborative environment by industry, government, and academia in an open systems collaborative environment
– Integration, verification, test, and release of Commercial-Off-The-Shelf (COTS) products for government use
– Rapidly prove operational utility of high technology solutions– Open systems and standards compliance evaluation,
documentation, and capabilities matrix
• Solve GAO identified big integrator problem1
– Organizing principle: “Give innovation a chance.”
• Consortium of large and small technology companies and facilities partners
1 Government Accounting Office-09-326SP, http://www.gao.gov/new.items/d09326sp.pdf
Infinite Dimensions
Solving Tomorrow’s Problems Today
CTE Network
6
Quality Technology Services (QTS) and Verizon (VZW) are the current facilities providers OCONUS Sites are VZW CONUS VZW: Engelwood, CO; Culpeper, VA, Miami, FL Remaining CONUS Sites are QTS, in particular the 1.3M sq. ft. Richmond, VA site
Infinite Dimensions
Solving Tomorrow’s Problems Today
Main CTE Experimentation Lab LocationRichmond- High Density Multi-Data Center Campus
500,000 Sq Ft of Planned Raised Floor Multiple Distinct Data Center Buildings
(Current Basis of Design)
1
3
2 Office Space
1.3 Million Sq Ft Campus
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Technical Challenges• Rapidly create complex, realistic, and scalable
networks of systems and component inter-relationships
• Distribution of autonomous controls and monitors• Implementation of complex webs of cause and effect• Dynamic alteration of the component execution
structure – Adaptation and evolution of the system
• Ability to handle billions of active processes in real-time– Harness power of sequential, distributed and/or parallel
processing – optimizing the use of any compute/network/storage configuration
– Smartphones to supercomputers
8
Infinite Dimensions
Solving Tomorrow’s Problems Today
Unmanned Systems Control
October 27, 2008 USIC Conference, San Diego, CA
9
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Hypersonic Anti-Ship Missile Challenges
• Critical Battlespace defense gap• Interactions happening faster than humans
can react• Current Command and Control (C2) is not
real-time, performance limited, and the only source of information
• Greatly affects real-time response for kinetic engagement capabilities
10
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Anti-Ship Missile Defense
Increase Response time for Fleet Simplify Installation, Maintenance, &
Operation
SBX
DDG
e.g. DF-21
UAS
Weapon
USV
Anti-Ship Missile
SpaceSensor
Infinite Dimensions
Solving Tomorrow’s Problems Today
12
Sea BasePerimeter
ASW Threat Identification Perimeter
SUW Threat Monitoring Perimeter
SUW Threat Identification Perimeter
18 nmi
15 nmi
50 nmi
15 nmi
ASMD Threat Perimeter
Infinite Dimensions
Solving Tomorrow’s Problems Today
13
Sea BasePerimeter
ASW Threat Identification Perimeter
SUW Threat Monitoring Perimeter
SUW Threat Identification Perimeter
ASMD Threat Perimeter
X
Infinite Dimensions
Solving Tomorrow’s Problems Today
Cloud of ProjectilesEach USV’s Gun Creates a Cloud
14
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
New Fleet Capabilities
• Automated Fleet Systems – Adaptive & composable– Knowledge built into systems– Emphasizes speed and flexibility
• Ability to fuse new information, knowledge, and structures rapidly– Mesh networks and systems– Example: Sensor net that automatically refocuses
based on accurate real time fused information
• C2 and fire control moves from Fleet platforms to Network
15
Infinite Dimensions
Solving Tomorrow’s Problems Today
Inter-system Interoperability and Interaction with Personnel “On Scene” Example: Nightingale II
16
Autonomous transit from starting point, to pick-up point, to medical unit
4
A. Call for MedEvac received at Nightingale Control
B. Best UAV is chosen automatically
C. Route is autonomously planned & uploaded
D. UAV is launched automatically
12
Autonomous collision & obstacle avoidance
Similar process for: Logistics, Combat Rescue, & Special Ops
No Fly Zone
C2
No Fly
Zone
A
B
C
D
1
Autonomous Clear-Zone
landing3
5
UGV+BEAR deploys
BEAR deploys
6 BEAR recovers &
Medic treats
7UAS/UGV/BEAR
system rejoins and goes to destination
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Nightingale II
• Autonomous:– VTOL UAV– High-mobility UGV(s)– Asset allocation and mission planning
• Interoperability / coordination with existing– Dismounted ground personnel– Air operations– Ground operations– Artillery & strike operations– Political & no-fly boundaries
– High-data-rate non-LOS communicationsOctober 27, 2008 USIC Conference, San Diego, CA 1
7
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Nightingale IIChallenges
• Autonomous VTOL UAV – Autonomous obstacle avoidance
• Wires, antennae, etc.– Sensor with sufficient resolution & range for
vehicle maneuverability limits– Day, night, weather, dust/sand/dirt (“brown
out”)
– Autonomous collision avoidance (other aircraft)• Small UAVs, birds, etc.
– Sensor with sufficient resolution & range for vehicle maneuverability limits
– Day, night, weather, dust/sand/dirt (“brown out”)
October 27, 2008 USIC Conference, San Diego, CA 18
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Nightingale IIChallenges
• Autonomous VTOL UAV (cont)– Autonomous LZ identification
• LZ size, geometry, roughness, slope• Ingress/egress flight-path
– Obstacles– Moving ground personnel, vehicles– Exposure to enemy
» Lines of fire» Exposure time
• Coordination with UGV mobility constraints– Navigable path between LZ and casualty
October 27, 2008 USIC Conference, San Diego, CA 19
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'1420
Nightingale IIChallenges
• Rough-terrain UGV(s) – Mobility
• Tracks, wheels, legs• Water crossing• Mud, sand, snow, ice• Interiors (stairs, doors, elevators, etc.)
– Autonomous capabilities • Beyond Grand Challenge, Urban Challenge
– Casualty extraction• Careful casualty handling • Does the UGV need a UGV?
October 27, 2008 USIC Conference, San Diego, CA
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
System/Component Integration
• Most well-known methods and technology originate in “business IT”…
• Optimizing execution and efficiency and enabling computation at scale is typically the province of “high performance computing”
• Real-time execution and synchronization are addressed by several communities, e.g. robotics
21
Infinite Dimensions
Solving Tomorrow’s Problems Today
Ease of Development/Integration/InteroperabilityImplement/Support Any Architecture
ApplicationConnectors
MessageBrokers
EnterpriseResource Managers
ApplicationServers
ComponentBrokers
Infinite Dimensions
Solving Tomorrow’s Problems Today
Data Source 1
Application 1
ComputeProcess 1
Compute Process M
Data Source 2
Data Source N
Application X
WAIT_FOR (DS1 semaphore, wait time1)
WAIT_FOR (DSP semaphore, wait timep)
Typical Complex Application-System
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
The DNA of Complex Systems
• Look at nature to understand complex systems
• Internal Processes• External Processes• Internal Events• External Events
• Intermix of all four is required– Implementing in a scalable manner is key
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Internal Processes Analogy: The Heart Beat
• Atria pump blood to ventricles, which contract
• Nonstop contractions are driven by the heart's electrical system
Internal Process: Synchronous or Asynchronous – Intrincsic Capabilities
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
External Processes Analogy: Pacemaker
• External process monitors and interacts with an object (i.e., a pacemaker monitors the heart’s rhythm)
• The electric current makes the heart beat within a certain range
External Process: Synchronous or Asynchronous – Monitor and Control
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Internal Events Analogy: Heart Attack
• Internal occurrence without pre-established time scale
• Certain factors cause the occurrence. Blood flow is restricted, or the nerve system, which controls the heart, malfunctions
Internal Occurrence: Irregular Time Scale – Intrinsic Capabilities
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
External Events Analogy: Defibrillation
• External event changes a passive object’s state (i.e., a defibrillator is used for resuscitation)
• External electrical shock is applied to the heart
• Foundational representation method
External Occurrence: Irregular Time Scale – Monitor and Control
Infinite Dimensions
Solving Tomorrow’s Problems Today
FE FE
FE FE FE
FE
FE
FE
FE FE
FE
FE FE FE FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
FE FE
User Application
Interface Services
Hardware Device
Interface Services
3D Visualization
Interface Services
Web Interface
Interface Services
CommunicationSpeed
Slow
Medium
Fast
Configurable Computational MeshPolymorphic Computing Architecture (PCA)
FE = Functional Element
29
Infinite Dimensions
Solving Tomorrow’s Problems Today
Component Repository
Composability Automation
CASE Tool EnvironmentUser Defined
IT SystemInterface
User DefinedHardwareInterface
Web ServicesAPI
(JNI, SOAP, OWL, etc.)
SharedMemory IP JTRS
ReflectiveMemory
Security State Saving Core Programming
DistributedObject Mgmt
Std App Dev Interface
SynchronizationManagement
Event ManagementServices
KnowledgeRepresentation
Integration Meta-DataData
Translation
Communication Services (Unicast, Multicast, Broadcast)
Common Application Services
Intelligent Application Services
System Execution Services
Service Decomposition
CompressionEncryption
BLOSLink-16 Others
30
Infinite Dimensions
Solving Tomorrow’s Problems Today
API ExampleTurret/Fire Control
Slew Elevate
Fire When Slew and Elevate are Complete
Process Firing Commands (and Queuing Them)
31
Infinite Dimensions
Solving Tomorrow’s Problems Today
Example: Turret Fire Commandvoid Turret::fire(){ P_VAR P_BEGIN(2)
// Wait until the turret movement is completed
WAIT_FOR(1, slewComplete, -1); WAIT_FOR(2, elevateComplete, -1);
// Fire the weapon, this would activate the real gun
Fire_M256(); RB_cout << "Flash, Boom, Bang, Echo" << endl; fireComplete = 1; P_END}
32
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Algorithm/Environment Integration
• Must be able to integrate arbitrary types of models (e.g. rule-based, network-based, fuzzy, combinations, etc.)
• A system must account for functionally disparate phenomena, in order to represent “intelligence” more effectively, for example:– Recognition of patterns– Adapting new solutions or strategies– Rule following
Infinite Dimensions
Solving Tomorrow’s Problems Today
Adaptive, Dynamic, Knowledge-based Execution/ControlConceptual Graphs
Computational Ontology Framework
ConceptsConcepts ActorsActorsRelationshipsRelationships
CAT STAT LOCSIT MATCAT STAT LOCSIT MATMAT
A Cat sits on a matA Cat sits on a mat Une Chat assis surune matteUne Chat assis surune matte
Infinite Dimensions
Solving Tomorrow’s Problems Today
Basic CG Formation Rules
Concept 1
Relationship 1
Relationship 1
Concept 2
Concepts or Other Actors
Actor 1
Actors or Relationships
Actors or Relationships
Concept 1
Concept
Concept or other Actors
Concept 1
Relationship 1
Actor 1
INVALID
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Extensions
• A CG execution engine with abilities to:– Embed in a hardware controller or a software program– Associate the software and hardware components as
indicated by the CG structure– Control, execute, and integrate the entire system
• Customization of concepts, relationships, and actors by the integrator, providing capability to: – Accommodate new hardware, or software
components, without returning to original developers for new versions of graph systems
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Extensions
• A mechanism to control the CG execution engine – Each node maintains a truth value to
managed graph execution– A concept from belief network theory
• Collectively, the extensions embodied in the system enable intelligent automation, execution, and control of complex hardware and software assemblies
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Summary of Extensions
• A collection of user-defined concept nodes• A collection of user-defined relationship nodes• A collection of user-defined actor nodes• A unique ID numbering scheme to identify every node, regardless of
type• A description of the connected nodes, and route of the connection• A list of references to input concept nodes (those with no incoming
arcs)– A valid CG must contain at least one input concept node
• A list of references to the output concept nodes (those with no outgoing arcs)– A valid CG must contain at least one output concept node
• A data structure that records the truth value of each node
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Paradigm of Execution
1. The system is starteda. Each component in the system is initialized
b. Synchronization relationships are established
c. Inputs are read and loaded into system components
2. The CG associated with the system is initialized and parseda. Each system component is associated with a CG element using the unique ID
tag (concept, relationship, actor)
b. Each component is registered in the correct collection mechanism for each type, using the unique ID tag
3. The system begins operating, with various evolution methods. For example:a. A user inputs information, which changes the state of the system (either event-
based, process-based, or simple update loop)
b. Regular update cycles occur for various system aspects
4. The system operation mechanism executes, and an Execution Cycle operation for the CG is activated
5. Each time a system operation mechanism is activated, step 4 is repeated
Infinite Dimensions
Solving Tomorrow’s Problems Today
CG ExecutionDividend: 9.0
Divisor: 4.0
Number: 144.0
Quotient: 2.0
Remainder: 1.0
SquareRoot: 12.0
Sum: 14.0
divide plus
sqrt
Dividend: 9
Divisor: 4
Number: 144
Quotient: 2
Remainder: 1
SquareRoot: 12
Sum: *s
divide plusplus
sqrt
IF ?r = 0IF ?r = 0
T
TT
T
TT
TT
T
F
Infinite Dimensions
Solving Tomorrow’s Problems Today
Example of a Path of Route in a CG
runs
Is a
generating
for
Responsible for
of Producing
AMachine
AnInterpreter
Algorithm
AnInterpretation
An Agent
A Graph A Behavior
Infinite Dimensions
Solving Tomorrow’s Problems Today
42
CSCI'14
Summary• Representation of complex webs of
synchronized causes and effects is central to the implementation of complex systems
• Computation, correlation of simultaneously evolving systems and interrelated phenomena
• Ability to control an activity based on a web of logic, and start another in response to dynamic conditions
• Achievement of scalability without loss of capability
Infinite Dimensions
Solving Tomorrow’s Problems Today
Backup
43
Infinite Dimensions
Solving Tomorrow’s Problems Today
CSCI'14
Software Development Framework: JEE
• Minimizes amount of code, people, skill-level, development, test, integration, and time required
• Standardizes processes, templates, and documentation
• Works with all modern languages and standard Application Programming Interfaces (APIs)
• Does not compromise any other software vendor’s intellectual property – typically a key cost, schedule, and performance driver
• Can be taught to all development team members in an extremely short period of time (< 1 to 2 days)
11
Infinite Dimensions
Solving Tomorrow’s Problems Today
Example: Building Automation Management System
Level 1Information
managementsystem
Level 2Information processingAnd supervisory system
Level 3Information processing/
Automation system
Level 4Process field
Infinite Dimensions
Solving Tomorrow’s Problems Today
Components in a BAMS
fanAir conditioning
PC
Heat energy
sanitary
lighting
Waste-managementelectricalacoustics
video
safety
elevator
Emergency powersupply
transformer
heating
Building
700