using feature locality can we
play

Using Feature Locality: Can We Motivation Leverage History to Avoid - PowerPoint PPT Presentation

Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Using Feature Locality: Can We Motivation Leverage History to Avoid Failures Failure Avoidance During Reconfiguration? Study Results


  1. Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Using Feature Locality: Can We Motivation Leverage History to Avoid Failures Failure Avoidance During Reconfiguration? Study Results Brady J. Garvin ∗ Myra B. Cohen Matthew B. Dwyer Conclusions and Future Work University of Nebraska–Lincoln September 4, 2011 Slide 1/52

  2. Outline Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation 1 Motivation Failure Avoidance Failure Avoidance 2 Study Results Study 3 Conclusions and Future Work Results 4 Conclusions and Future Work 5 Slide 2/52

  3. Spirit Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study Results Conclusions and Future Work Adirondack, Spirit’s first rock, whose investigation would be interrupted by a bout of insomnia, fever, and delirium. Slide 3/52

  4. What we Know from Testing Using Feature Locality: Can We Leverage History Highly-Configurable Systems to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Spirit, a highly-configurable system, was exhibiting a Study configuration-dependent failure [Adler 2004]. Results Highly-configurable systems tend to do this [Kuhn et al. 2004,Yilmaz et al. 2006,Qu et al. 2008]. Conclusions and Future Work Configuration-dependence makes failures harder to find in testing, but easier to work around. Slide 4/52

  5. What we Know from Designing Using Feature Locality: Can We Leverage History Autonomous Systems to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Adaptations, including online reconfiguration, affect Failure Avoidance reliability. Study Reliability can be predicted fairly well from a few key Results parameters [Ahmed et al. 2010]. Conclusions and But until the system has been running for a while under Future Work various configurations, we will have observed only a handful of discrete failures, which makes it difficult to build a reliability model and identify those parameters. Slide 5/52

  6. What we Know from Fault Prediction Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Faults exhibit many forms of locality, both temporal and Study spacial [Hassan et al. 2005,Kim et al. 2007]. Results In consequence, failures also show locality. Conclusions and Future Work Slide 6/52

  7. Hypothesis Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Reconfiguration workarounds effective against one Study failure tend to be effective against others. Results We call this phenomenon feature locality of failures. Conclusions and Future Work Slide 7/52

  8. Outline Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation 1 Motivation Failure Avoidance Failure Avoidance 2 Study Results Study 3 Conclusions and Future Work Results 4 Conclusions and Future Work 5 Slide 8/52

  9. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Navigation IDD . . . Blind Stowed . . . Study Hazard Avoidance APXS . . . Results Autonomous MB . . . Conclusions and Wheel Wiggle RAT . . . Future Work Slide 9/52

  10. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Navigation IDD . . . Blind Stowed . . . Study Hazard Avoidance APXS . . . Results Autonomous MB . . . Conclusions and Wheel Wiggle RAT . . . Future Work Slide 10/52

  11. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study Results Conclusions and Future Work Navigation Slide 11/52

  12. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Navigation IDD . . . Blind Stowed . . . Study Hazard Avoidance APXS . . . Results Autonomous MB . . . Conclusions and Wheel Wiggle RAT . . . Future Work Slide 12/52

  13. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 13/52

  14. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Navigation IDD . . . Blind Stowed . . . Study Hazard Avoidance APXS . . . Results Autonomous MB . . . Conclusions and Wheel Wiggle RAT . . . Future Work Slide 14/52

  15. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 15/52

  16. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 16/52

  17. Configuration Space Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 17/52

  18. Algorithms Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance 1 Analyze failures Study 2 Guard against dangerous configurations Results 3 Workaround new failures Conclusions and Future Work Slide 18/52

  19. Algorithm 1: Analyze Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 19/52

  20. Algorithm 1: Analyze Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 20/52

  21. Algorithm 1: Analyze Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 21/52

  22. Algorithm 1: Analyze Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 22/52

  23. Algorithm 1: Analyze Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 23/52

  24. Algorithm 2: Guard Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 24/52

  25. Algorithm 2: Guard Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 25/52

  26. Algorithm 3: Reconfigure Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 26/52

  27. Algorithm 3: Reconfigure Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 27/52

  28. Algorithm 3: Reconfigure Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 28/52

  29. Algorithm 3: Reconfigure Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 29/52

  30. Algorithm 3: Reconfigure Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation Failure Avoidance Study IDD Results Conclusions and Future Work Navigation Slide 30/52

  31. Outline Using Feature Locality: Can We Leverage History to Avoid Failures During Reconfiguration? ASAS 2011 Motivation 1 Motivation Failure Avoidance Failure Avoidance 2 Study Results Study 3 Conclusions and Future Work Results 4 Conclusions and Future Work 5 Slide 31/52

Recommend


More recommend