distributed w atchpoints debugging very large ensem bles
play

Distributed W atchpoints: Debugging Very Large Ensem bles of Robots - PowerPoint PPT Presentation

Distributed W atchpoints: Debugging Very Large Ensem bles of Robots De Rosa, Goldstein, Lee, Campbell, Pillai Aug 19, 2006 Motivation Distributed errors are hard to find with traditional debugging tools Centralized snapshot algorithms


  1. Distributed W atchpoints: Debugging Very Large Ensem bles of Robots De Rosa, Goldstein, Lee, Campbell, Pillai Aug 19, 2006

  2. Motivation • Distributed errors are hard to find with traditional debugging tools • Centralized snapshot algorithms – Expensive – Geared towards detecting one error at a time • Special-purpose debugging code is difficult to write, may itself contain errors 2 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  3. Expressing and Detecting Distributed Conditions “How can we represent, detect, and trigger on distributed conditions in very large multi-robot systems?” • Generic detection framework, well suited to debugging • Detect conditions that are not observable via the local state of one robot • Support algorithm-level debugging (not code/ HW debugging) • Trigger arbitrary actions when condition is met • Asynchronous, bandwidth/ CPU-limited systems 3 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  4. Distributed/ Parallel Debugging: State of the Art Modes: • Parallel: powerful nodes, regular (static) topology, shared memory • Distributed: weak, mobile nodes Tools: • GDB • printf() • Race detectors • Declarative network systems with debugging support (ala P2) 4 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  5. Exam ple Errors: Leader Election Scenario: One Leader Per Tw o-Hop Radius 5 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  6. Exam ple Errors: Token Passing Scenario: I f a node has the token, exactly one of it’s neighbors m ust have had it last tim estep 6 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  7. Exam ple Errors: Gradient Field Scenario: Gradient Values Must Be Sm ooth 7 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  8. Expressing Distributed Error Conditions Requirements: • Ability to specify shape of trigger groups • Temporal operators • Simple syntax (reduce programmer effort/ learning curve) A Solution: • Inspired by Linear Temporal Logic (LTL) – A simple extension to first-order logic – Proven technique for single-robot debugging [ Lamine01] • Assumption: Trigger groups must be connected – For practical/ efficiency reasons 8 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  9. W atchpoint Prim itives nodes(a,b,c); n(b,c) & (a.var > b.var) & (c.prev.var != 2) • Modules (implicitly quantified over all connected sub-ensembles) • Topological restrictions (pairwise neighbor relations) • Boolean connectives • State variable comparisons (distributed) • Temporal operators 9 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  10. Distributed Errors: Exam ple W atchpoints nodes( a,b,c) ;n( a.b) & n( b,c) & ( a.isLeader = = 1 ) & ( c.isLeader = = 1 ) nodes( a,b,c) ;n( a,b) & n( a,c) & ( a.token = = 1 ) & ( b.prev.token = = 1 ) & ( c.prev.token = = 1 ) nodes( a,b) ;( a.state - b.state > 1 ) 1 0 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  11. W atchpoint Execution 1 nodes(a,b,c)… 2 3 2 1 1 9 9 2 1 1 2 3 4 5 6 7 8 � 1 9 10 9 10 11 12 13 14 15 16 . 17 18 19 20 21 22 23 24 . . 25 26 27 28 29 30 31 32 . 1 1 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  12. Perform ance: W atchpoint Size • 1000 modules, running for 100 timesteps • Simulator overhead excluded • Application: data aggregation with landmark routing • Watchpoint: are the first and last robots in the watchpoint in the same state? 1 2 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  13. Perform ance: Num ber of Matchers • This particular watchpoint never terminates early • Number of matchers increases exponentially • Time per matcher remains within factor of 2 • Details of the watchpoint expression more important than size 1 3 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  14. Perform ance: Periodically Running W atchpoints 1 4 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  15. Future W ork • Distributed implementation • More optimization • User validation • Additional predicates 1 5 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  16. Conclusions • Simple, yet highly descriptive syntax • Able to detect errors missed by more conventional techniques • Low simulation overhead 1 6 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  17. Thank You

  18. Backup Slides 1 8 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

  19. Optim izations • Temporal span • Early termination • Neighbor culling • (one slide per) 1 9 8 / 1 9 / 2 0 0 6 Distributed W atchpoints

Recommend


More recommend