Video of the Day LA Express Park Explained 1 Proposal One - PowerPoint PPT Presentation

Video of the Day Ø LA Express Park Explained � 1

Proposal Ø One proposal/team, 1 page � q Team members � q Concise description of project � q Responsibilities of each member � q Specific equipment needed � Ø Written proposal due: 9/20 th , 11:59pm � q Email to Mo � 2

Proposal Presentation Ø 8 min per group � q Including 3 min for questions � Ø Your elevator pitch! � Ø Email Mo your slides by 11am on 9/20 th � q All use Mo’s laptop à reduce context switches � 3

Agilla: Mobile Agent Middleware for Wireless Sensor Networks Che Cheny nyang ang Lu Lu � Computer Science and Engineering �

Motivation Ø Adaptive applications in wireless sensor networks � Ø Adapt to changes in � q the environment � q the network � q user requirements � 5

Example: Forrest Ø Three applications: 1) Environmental Monitoring, 
 2) Fire Detection, 3) Fire Tracking � 6

Example: Forrest with Agilla Ø Sensor network as a shared computing resource � Flexible application deployment � q Env. monitoring agent Fire detection agent Fire tracking agent 7

Example: Cargo Tracking Thousands of containers leave/join network per day � Ø Software need to be changed on the fly due to � Ø Departure and arrival of containers � q Container’s country and company � q Change in security levels and policies � q Change in tracking technologies � q Ø Agilla Agilla: support adaptive adaptive applications in wireless sensor networks � 8

Agilla’s System Architecture Node @ (2,1) Node @ (1,1) Agents Agents migrate remote Neighbor Neighbor access Tuplespace Tuplespace List List Middleware Services Middleware Services Agilla Middleware Agilla Middleware TinyOS TinyOS 9

Agilla’s Computational Model PC Clone or Stack Migrate Code Condition Codes Heap Two variants of each: 1) Strong (code + state) 2) Weak (code only) 10

Mobile Agent Model Ø Mobile agent: unit of execution that can move across hosts � Ø Multiple mobile agents may share a host � Ø Spatiotemporal decoupling => need shared data space with uniform interfaces � 11

Tuple Space-Based Coordination Content-addressable shared memory � Ø Tuple – A set of data fields � q Template – A pattern that matches particular tuples � q Provides spatiotemporal decou oupling in unreliable networks � Ø “rout” “in” “in” “out” Tuplespace Tuplespace 12

Tuple Space Model Tuple: collection of data � Ø <15, “str”, false> � q Tuple space: shared data space containing tuples � Ø Template: patterns which match tuples � Ø <15, *, false> matches <15, “str”, false> � q Common tuple space operations: � Ø ou out(t): add tuple t to tuple space � q rd(T) rd(T): copy tuple matching template T from tuple space � q in in(T) (T): remove tuple matching template T from tuple space � q re regrxn(T, c): grxn(T, c): execute code segment C when tuple matching q template T is placed in tuple space � 13

Agilla Tuple Space API Remotely accessible localized tuple spaces � Ø Stores context information � out in Ø Facilitates inter-agent communication � Ø Tuplespace Local Remote out : � � insert � in : � � � remove � rout : � insert � rd : � � � read � rinp : � probing remove � inp : � � probing remove � rrdp : � probing read � rdp : � � probing read � rrdpg : � probing group read (1-hop) � regrxn : � � register reaction � deregrxn : � deregister reaction � 14

Location-Base Addressing Ø Nodes are addressed by location � (3,3) (3,1) (3,2) (2,2) clone to (3,3) clone to (3,1) Fire Detection (1,3) (1,1) Agent 15

Implementation on TinyOS 1.1.13 Agilla is available for Mica2 and MicaZ motes � Ø 4 agents/node � q Agent Injector � Ø Written in Java � q Remote Injection via RMI � q Key Challenges � Ø Network bandwidth � q • Compact instructions � Memory � q • ROM: 54.7KB of 128KB � • RAM: 3.5KB of 4KB � Message loss � q • Agent-level ARQ � 16

Experimental Set-up 6x9 Mica2 Mote 
 Ø Testbed � Multi-hop Grid � Ø One base station � Ø 17

Performance Evaluation: migration vs. remote tuple space access Migration instructions are more reliable because of hop-by-hop acknowledgements… …but remote tuplespace operations have less overhead 18

Agilla Instruction Execution Times Local Operations Remote Operations 19

Experiences Ø Fire Detection & Tracking � Presented at IPSN 2005 � q Ø Intruder Detection and Tracking � Agents guard network perimeter and follow intruders � q Periodically report intruder location to base station � q Ø Cargo & Inventory Management � In collaboration with Boeing � q Mobile agents load manifests from RFID, find items, q detect security breaches, and send alert to Internet gateways. � Demo at SenSys 2005 � q 20

Fire Tracking Video Video available at: http://mobilab.wustl.edu/projects/agilla 21

Related Work Ø Distributing inanimate code modules � XNP [xbow’03], Deluge [sensys’04], MNP [icdcs’05], 
 q SOS [mobisys’05] � Maté/Bombilla [asplos’02] � q Ø Mobile Agent-Like Middleware � Sensorware [mobisys’03] � q • Weak migration only � Smart Messages [Kang‘04] � q • No remote interactions � • Single thread per node � 22

Agilla Summary Ø Mobile agent middleware simplifies application deployment & increases network flexibility � Ø Agilla middleware services � Agent mobility � q Tuple space-based coordination � q Location-centric addressing � q Context discovery � q Ø Empirical results demonstrate Agilla is re reliable liable and effic efficien ient � 23

References C.-L. Fok, G.-C. Roman and C. Lu, Agilla: A Mobile Agent Middleware for Self-Adaptive Wireless Sensor Networks, ACM Transactions on Autonomous and Adaptive Systems, Special Issue on Self-Adaptive and Self-Organising Wireless Networking Systems, 4(3), Article 16, July 2009. Agilla: http://mobilab.wustl.edu/projects/agilla • Source Code • Documentation • Tutorials • Experience Reports 24

Agimone: Middleware Integration of Sensor and IP Networks � � � Chenyang Lu � Computer Science and Engineering �

Motivating Application: Cargo Tracking Ship Train Sensor Sensor Network Network IP Network Sensor Sensor Network Network Truck Customer, Shipper, DHS, CBP Shipping Yard 26

Problem Statement Ø Weak radio => low radio range => isolated networks => reduced utility � Ø Communication restrictions can be eased by connecting WSNs using a common IP network � Ø Custom application-specific software is written to connect sensor and IP networks � Sensor and IP networks have vastly different q characteristics and capabilities � Not reusable, error prone, inflexible � q 27

Background: Two Distinct Middlewares Node (1,1) Node (2,1) Mobile Host 1 Mobile Host 2 Agents Agents Agents Agents migrate migrate AQL AQL AQL AQL AQL AQL AQL AQL remote access TS TS TS TS TS TS TS TS remote access Neighbors Neighbors Neighbors Neighbors Tuple Space Tuple Space Limone Server Limone Server Agilla Middleware Agilla Middleware TinyOS TinyOS Java VM Java VM MICA2 Mote MICA2 Mote Agilla : Sensor Network Middleware Limone : IP Network Middleware • Mobile Agents • Mobile Agents • Host-level tuple spaces • Agent-level tuple spaces • Host-level neighbor list • Agent-level neighbor list Severe resource limitations Resource rich 28

Programming Models Agilla Limone Ø Agilla byte code � Ø Java � Ø Hosted on custom VM � Ø Hosted on Java VM � Ø Handful of well-known Ø Rich data types � data types � 29

Agimone Network Architecture System Components 30

Agimone Services Shared data space � Ø AgimoneAgent has own Agilla TS � q Agilla agents access TS directly � q AgimoneAgent wraps Agilla TS in q Limone TS API � Sensor network discovery � Ø AgimoneAgent advertises self with q Limone Registry � Limone Registry distributes ad to q other AgimoneAgents � Ads stored in AgimoneAgent’s q Agilla tuple space � 31

Agimone Services Inter-network Migration Multi-step process � Ø Though internally Ø complex, appears simple to Agilla agent developer � 32

Cargo Tracking Revisited Watchdog Agents � Ø Monitors sensors � q Sends alert to base station � q Port authority’s Limone agent q reacts to it � 17 lines of Agilla code, 11 q lines of Java code � Intrusion Search Agent � Ø Searches multiple WSNs for q alert tuples � • Autonomously crosses WSN boundaries � 23 lines of Agilla code, no q additional Java code � 33

Experimental Setup 100Mbps Ethernet IBM R40 Laptop • 1.5GHz Pentium M • 512MB RAM • Windows XP Mica2 Mote on Gateway • 7.4MHz Atmel 128L 115kbps serial • 4KB RAM • TinyOS 34