Complex Environment Evolution

Complex Environment Evolution

Challenges with Semantic Service Infrastructures

- Andrej Eisfeld- Achim P. Karduck- David McMeekin IEEE DEST: 18 - 20 June 2012

Complex Environment Evolution2

2

Structure

Background

Semantic Agents

Evaluation

Conclusion

Complex Environment Evolution3

3

Background Semantic Agents Use Case Conclusion

Smart Camp

Aim: Reduce energy consumption in camps

Example:

Energy costs: 2.000.000 AUD / year

25% savings potential

Main Smart Camp System components:

Smart Home Controller (SHC)

Smart Camp Management Unit (SCMU)

Complex Environment Evolution4

4

Background Semantic Agents Use Case Conclusion

Problem I

Continuing Change

“E-type systems must be continually adapted or they become progressively less satisfactory”

Continuing Growth

“The functional content of E-type systems must be continually increased to maintain user satisfaction over

their lifetime”

Complex Environment Evolution5

5

Background Semantic Agents Use Case Conclusion

Problem II

Multiple software systems in service infrastructure

Evolution more difficult due to dependencies

Complex Environment Evolution6

6

Background Semantic Agents Use Case Conclusion

Semantic Service Approaches

Approach Loose Coupling

WSDL2.0 + SAWSDL x

HTML + SA-REST

HTML + hRESTs + MicroWSMO

EXPRESS

ReLL

JSON-LD

Comparison of multiple Semantic Service aproaches

Complex Environment Evolution7

7

Background Semantic Agents Use Case Conclusion

Linked Data II

JSON-LD is resource orientated

Linked Resources Graph (LRG):

Complex Environment Evolution8

8

Background Semantic Agents Use Case Conclusion

Idea I : LRG Ontology

Resource Discovery

Resource Composition

Resource Invocation

Complex Environment Evolution9

9

Background Semantic Agents Use Case Conclusion

Idea II : Ontology Paths

Permitted Ontology Path (POP)

Not Permitted Ontology Path (NPOP)

POP + NPOP → Restrictions for LRG traversal

Complex Environment Evolution10

10

Background Semantic Agents Use Case Conclusion

Semantic Handler

Semantic Request Handler

Resorce Discovery + Composition + Invocation

Semantic Response Handler

Data Discovery + Dynamic Code Reuse

Complex Environment Evolution11

11

Background Semantic Agents Use Case Conclusion

Agent Communication

1) Define Goal

2) Traverse LRG

3) Retrieve Response

4) Process Response

Complex Environment Evolution12

12

Background Semantic Agents Use Case Conclusion

A Semantic Camp

SCMU and SHCs as Semantic Agents

Flexibility for Resource's location and content

Functionality enrichment without recompilation

Complex Environment Evolution13

13

Background Semantic Agents Use Case Conclusion

Setting

Linked Resources GraphSmart Camp Ontology

Complex Environment Evolution14

14

Background Semantic Agents Use Case Conclusion

Resource Discovery

Smart Camp Ontology Linked Resources Graph

Complex Environment Evolution15

15

Background Semantic Agents Use Case Conclusion

Representations

{

"@context":{

"onto":"http://www.smartcamp.org/onto"

"door":"onto#DoorSensor"

"value":"onto#sensorValue"

},

"@type":"door",

"value":true

}

{

"@context":{

"onto":"http://www.smartcamp.org/onto"

"motion":"onto#MotionSensor"

"value":"onto#sensorValue"

},

"@type":"motion",

"valueZ":false

}

Complex Environment Evolution16

16

Background Semantic Agents Use Case Conclusion

Composed Representation

{

"@context":{

"motion":"http://www.smartcamp.org/ontology#MotionSensor",

"door":"http://www.smartcamp.org/ontology#DoorSensor",

"value":"http://www.smartcamp.org/ontology#sensorValue"

},

"@type":"http://www.smartcamp.org/ontology#Sensor",

"motion":{

"value":false

},

"door":{

"value":true

}

}

Complex Environment Evolution17

17

Background Semantic Agents Use Case Conclusion

What if ...

● Requirements change → new sensors● Requirements change → obsolete sensors

Complex Environment Evolution18

18

Background Semantic Agents Use Case Conclusion

Summary

Chosen technologies: JSON-LD + OWL

Model of a Semantic Agent

Higher evolvability in evolution scenario

Ontology Evolution may reduce assessed evolvability

Complex Environment Evolution19

19

Background Semantic Agents Use Case Conclusion

Outlook

Implementation

Research Ontology Evolution & Versioning

Service Discovery in a Smart City

Complex Environment Evolution21

21

References● M. Lehman. On understanding laws, evolution, and conservation in the large-

program life cycle. Journal of Systems and Software, 1:213–221, 1980● H. P. Breivold, I. Crnkovic, R. Land, and S. Larsson. Using dependency model

to support software architecture evolution. In Automated Software Engineering - Workshops, 2008. ASE Workshops 2008. 23rd IEEE/ACM International Conference on, pages 82–91, 2008.

● P.V.D. Laar and T. Punter. Views on Evolvability of Embedded Systems. Springer, 2010.

● Ora Lassila, Tim Berners-Lee, James A. Hendler. The semantic web. Scientific American, 284(5):34–43, 2001.

● http://www.cs.helsinki.fi/research/roosa/images/serious-logo-final.jpg● http://applicanttracking.files.wordpress.com/2010/06/evolution.jpg● http://informatique.umons.ac.be/genlog/images/wordle.jpg● http://www.johnbendever.com/wp-content/uploads/question.jpg

Complex Environment Evolution22

22

DNS Service Discovery

Different types of resource records

PTR: Defines references to other domains

SRV: Defines a service location

TXT: Used to add meta-data

------------------------------------------------------------------

General usage:

serviceType PTR serviceInstance

serviceInstance SRV serviceLocation

TXT serviceMetaData