Diagnosing and Repairing Data Anomalies in Process Models

1,599
-1

Published on

Workshop presentation given by Niels Lohmann on September 7, 2009 in Ulm, Germany at the 5th International Workshop on Business Process Design (BPD 2009); part of BPM 2009.

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
1,599
On Slideshare
0
From Embeds
0
Number of Embeds
6
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Diagnosing and Repairing Data Anomalies in Process Models

  1. 1. Diagnosing and RepairingData Anomalies in Process ModelsAhmed AwadHassoPlattner Institute, Potsdam, GermanyGero Decker HassoPlattner Institute, Potsdam, GermanyNielsLohmann University of Rostock, Germany<br />
  2. 2. Correctness of Process Models<br />widely accepted: soundness<br />no deadlocks<br />no livelocks<br />proper termination<br />no dead activities<br />These are control flow aspects!<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  3. 3. Example Process: Insurance Claim Handling<br />sound: every claim will eventually be closed<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  4. 4. Data in BPMN<br />Data objects<br />Data states (no explosion)<br />Object life cycles /control flow refinement<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  5. 5. Process Model with Data<br />This model contains five deadlocks!<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  6. 6. Classes of Errors<br />too restrictive preconditions (close and fraudulent claims)<br />implicit routing (XOR vs. fraud evaluation)<br />implicit execution order (pay vs. file)<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  7. 7. Outline<br />control flow + data flow = interesting problems ✔<br />formalization BPMN’s data aspects<br />detection, diagnosing, and repairing of data anomalies<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  8. 8. 2<br />FormalizationBPMN’s dataaspects<br />
  9. 9. BPMN and Petri nets<br />BPMN:<br />a graphical notion<br />support of concurrency<br />Petri nets:<br />a graphical notion<br />support of concurrency<br />formal foundation<br />broad tool support<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />Dijkman et al. definePetri net semanticsfor BPMN’s control flow <br />
  10. 10. Petri net formalization (control flow)<br />pattern-based translation<br />complete example (control flow):<br />analysis tools can check soundness<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  11. 11. BPMN and Petri nets (2)<br />BPMN:<br />a graphical notion<br />support of concurrency<br />several aspects in one model<br />Petri nets:<br />a graphical notion<br />support of concurrency<br />formal foundation<br />broad tool support<br />simple composition notions<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  12. 12. Formalization of BPMN’s data objects<br />changing a state<br />reading a state<br />changing to several possible states<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  13. 13. Data flow models<br />data flow models for settlement and claim data object<br />control flow model<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  14. 14. Bringing it all together<br />synchronization of data flow and control flow by transition fusion<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  15. 15. 3<br />Detection,diagnosing,and repair ofdata anomalies<br />
  16. 16. Detection of data anomalies<br />standard soundness checker (Woflan, LoLA) will find deadlocks<br />provides counterexample (= trace)<br />does not differentiate data flow and control flow<br />gives no diagnosis/repair information<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  17. 17. Diagnosis data anomalies<br />exploit information on model:<br />control flow is sound<br />place models either control flow or data flow<br />each data object can only be in one state<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  18. 18. Diagnosing and fixing too restrictive preconditions<br />Problem: if data is set to [a], activity B is disabled<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  19. 19. Diagnosing and fixing too restrictive preconditions<br />control flow is sound<br />deadlock in composite model: missing data tokens<br />for each deadlock:<br />determine missing data tokens<br />change model<br />data tokens are present, or<br />drop data dependency<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  20. 20. Diagnosing and fixing implicit routing<br />decision [b] vs. [c] has to be synchronized with XOR-split<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  21. 21. Diagnosing and fixing implicit routing<br />partition state space with respect to data states<br />if a decision inside a partition leads to a deadlock, this decision is “unsynchronized”<br />synchronize decisions according to data<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  22. 22. Diagnosing and fixing implicit execution order<br />transitions A and B are in concurrent the control flow model,but share data place<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  23. 23. Take home points<br />data objects can introduce errors to a model<br />Petri nets allow for compositional models of data and control flow<br />data anomalies can be detected, diagnosed,and (sometimes) automatically fixed<br />Future work:<br />automated mapping back to the BPMN model<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />
  24. 24. Final slide<br />Thank you for your attention!<br />Your todos:<br />discuss with me<br />talk to Ahmed and Gero<br />attend the soundness talk(Thursday, after the keynote)<br />get the slides athttp://slideshare.net/correctsystems<br />Diagnosing and Repairing Data Anomalies in Process Models<br />07.09.09<br />

×