1 / 7

A Dimensionless Graceful Degradation Metric for Quantifying Resilience

A Dimensionless Graceful Degradation Metric for Quantifying Resilience. Jacob Beal. Eval4SASO Workshop IEEE SASO September, 2012. Problem: What is “Resilience”?. Which of these two systems is better? Prior work has few quantitative metrics, and none are comparable and feasible to compute.

reba
Download Presentation

A Dimensionless Graceful Degradation Metric for Quantifying Resilience

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A Dimensionless Graceful Degradation Metric for Quantifying Resilience Jacob Beal Eval4SASO Workshop IEEE SASO September, 2012

  2. Problem: What is “Resilience”? Which of these two systems is better? • Prior work has few quantitative metrics, and none are comparable and feasible to compute

  3. Formalizing Graceful Degradation • Domain of possible configurations: • Metrics of system performance: • Interpretation as acceptable/degraded/failing: • Graceful Degradation Metric

  4. Example: Failing Failing Acceptable Acceptable Axis of least graceful degradation Degraded Axis of least graceful degradation Degraded F D F D A D A D F F More graceful

  5. Computing Graceful Degradation • Analytic solutions generally impossible • Thorough surveys infeasible for most systems • Orthogonal (1-parameter) perturbation surveys: Example from MADV study of functional blueprints: [Adler et al., 2011] G(D) = 0.2

  6. Handling Parameter Dependencies • Orthogonal perturbation assumes quasi-linearity Often not true! • Alternate approach: random perturbation vectors • Based on manifold learning in [Freund et al. 2007]

  7. Contributions • Proposed metric for graceful degradation • Readily comparable: dimensionless, invariant to linear transformations and extra dimensions • Feasible to compute via perturbation surveys • Proposed random perturbation survey method

More Related