Correcting Deadlocking Service Choreographies Using a Simulation-Based Graph Edit Distance
2 likes794 views
The document discusses a method for correcting choreographies in service orchestration by using graph similarities to avoid deadlocks. It highlights the limitations of synthesizing services independently from a scapegoat participant and introduces a simulation-based approach to find more suitable alternatives without exhaustive checking. The approach is implemented in a tool named Rachel, which helps in editing choreographies for real-life processes, while also raising questions about the complexity and ability to support more complex edits.
![State-of-the-art choreography analysis 3translate BPEL choreography into formal modelcheck for deadlocksif deadlock found:choose a "scapegoat" participantremove it from the choreographysynthesize a corrected version (if possible)retranslate synthesized participant back to BPELFull tool support available! [WS-FM2007]](https://image.slidesharecdn.com/08-09-02correctingchoreographiesusinggraphsimilarities-090902053909-phpapp01/85/Correcting-Deadlocking-Service-Choreographies-Using-a-Simulation-Based-Graph-Edit-Distance-3-320.jpg)
![Operating Guidelines in a Nutshell9automaton annotated with Boolean formulaecharacterizes all partners of a service [ATPN2007]partner iffsimulated by OG + fulfilling the annotations](https://image.slidesharecdn.com/08-09-02correctingchoreographiesusinggraphsimilarities-090902053909-phpapp01/85/Correcting-Deadlocking-Service-Choreographies-Using-a-Simulation-Based-Graph-Edit-Distance-9-320.jpg)
![Simulation-based graph similarityIdea: find a similarity that respects simulationcompare two states and find best transitionsw.r.t. successor states [TACAS2006]aabccbddd14](https://image.slidesharecdn.com/08-09-02correctingchoreographiesusinggraphsimilarities-090902053909-phpapp01/85/Correcting-Deadlocking-Service-Choreographies-Using-a-Simulation-Based-Graph-Edit-Distance-14-320.jpg)
![22[TACAS2006] Oleg Sokolsky, SampathKannan, and Insup Lee. Simulation-based graph similarity. In TACAS 2006, volume 3920 of LNCS, pages 426–440. Springer, 2006.[WS-FM2007] Niels Lohmann, Oliver Kopp, Frank Leymann, and Wolfgang Reisig. Analyzing BPEL4Chor: Verification and participant synthesis. In WS-FM 2007, volume 4937 of LNCS, pages 46–60. Springer, 2008.[ATPN2007] Niels Lohmann, Peter Massuthe, and Karsten Wolf. Operating guidelines for finite-state services. In ICATPN 2007, volume 4546 of LNCS, pages 321-341. Springer, 2007.References](https://image.slidesharecdn.com/08-09-02correctingchoreographiesusinggraphsimilarities-090902053909-phpapp01/85/Correcting-Deadlocking-Service-Choreographies-Using-a-Simulation-Based-Graph-Edit-Distance-22-320.jpg)
Correcting Deadlocking Service Choreographies Using a Simulation-Based Graph Edit Distance
- 1. Correcting Choreographiesusing Graph SimilaritiesNiels LohmannBPM 2008 ▪ Milan ▪ 2 September 2008http://service-technology.org/bpm2008UNIVERSITÄT ROSTOCK
- 2. 2a deadlock in a choreographyHow can weavoid this?What wentwrong?Who is toblame?http://thisboyissmooth.wordpress.com/2008/02/12/sistemas-operativos-e-deadlocks/
- 3. State-of-the-art choreography analysis 3translate BPEL choreography into formal modelcheck for deadlocksif deadlock found:choose a "scapegoat" participantremove it from the choreographysynthesize a corrected version (if possible)retranslate synthesized participant back to BPELFull tool support available! [WS-FM2007]
- 4. Problem with that approach4the synthesized serviceis built independently from the scapegoatgives no information what was changedis correct, yet might not cover the original intention
- 7. Fix the customer service7✗scapegoat✓✓better choice
- 8. Setting8given:a service (the scapegoat)a set ofcandidatesfind:the candidate that is most similar to the scapegoat…without sequentially checking all candidates
- 9. Operating Guidelines in a Nutshell9automaton annotated with Boolean formulaecharacterizes all partners of a service [ATPN2007]partner iffsimulated by OG + fulfilling the annotations
- 10. OG characterizes all possible corrections10and 2001 additional candidates✓✗✓
- 11. Setting (refined)11given:a service automaton (the scapegoat)an operating guideline characterizing allcandidatesfind:the candidate that is most similar to the scapegoat…without sequentially checking all candidates
- 12. Graph edit distancea measure to compare graphs: edit distance= no. of needed actions to achieve graph isomorphismmaybe associated with a cost functionbamodify b to aad0,5adadd c branchecdcde0,7delete e branche0,312
- 13. Graph edit distance vs. behavior!a?b!c?b!alow edit distanceunsimilar behaviorhigh edit distanceequivalent behavior!c!a?b?b!ahigh edit distanceunsimilar behavior!c13
- 14. Simulation-based graph similarityIdea: find a similarity that respects simulationcompare two states and find best transitionsw.r.t. successor states [TACAS2006]aabccbddd14
- 15. Simulation-based graph similarityMismatches are treated with stuttering stepspenalize stuttering by label similarity functionchoose best pairs by maximizing label similarityaabccbedεf15
- 16. Simulation-based edit distancelabel similarity function defines an edit distance(a,a) ➙keep a(e,d) ➙change e to d(ε,x) ➙insert x(f,ε) ➙delete fvalues can be derived from semantic Webinformation(!€,!$) or (?receipt,?confirmation) rather high(!login,?invoice) rather lowaabcbcedεf16
- 17. Simulation and OG matching17Simulation is only one part of the OG matchingnext step: make edit distance aware of formulas✗
- 18. Respect formulas18instead of comparison with the OG's structure…compare with satisfying assignments of the formulaworst-case complexity: O(|QSA|⋅|QOG|⋅2|I|⋅|I|!)assignmentsedge permutations
- 19. Experimental results19Simulation- and matching-based edit distance implemented in tool RAChELDynamic programming exploits structure of the problem Most results within few secondsExceptions have near-worst-case structure/formulasRepairing Automata forChoreographies by Editing Labels
- 20. 20edit actions can be mapped back to original serviceresult can be influenced by adjusting label similarities Fixing the example with Rachel
- 21. 21choreographies can be fixed using the edit distancecan help to only change little partsof the scapegoatprototype shows that fixing of real-life processesworks(tool + slides at http://service-technology.org/bpm2008)Open questions:Which service to fix?What about cyclic or nondeterministic services?How does the mapping backto BPEL really work?Can we support more elaborate edit actions?Can heuristics help to improve performance? Take home pointsThank you!Any questions?
- 22. 22[TACAS2006] Oleg Sokolsky, SampathKannan, and Insup Lee. Simulation-based graph similarity. In TACAS 2006, volume 3920 of LNCS, pages 426–440. Springer, 2006.[WS-FM2007] Niels Lohmann, Oliver Kopp, Frank Leymann, and Wolfgang Reisig. Analyzing BPEL4Chor: Verification and participant synthesis. In WS-FM 2007, volume 4937 of LNCS, pages 46–60. Springer, 2008.[ATPN2007] Niels Lohmann, Peter Massuthe, and Karsten Wolf. Operating guidelines for finite-state services. In ICATPN 2007, volume 4546 of LNCS, pages 321-341. Springer, 2007.References