Runtime verification in context: can optimizing error detection improve fault diagnosis?

  • Authors:

  • Matthew B. Dwyer;Rahul Purandare;Suzette Person

  • Affiliations:

  • Department of Computer Science and Engineering, University of Nebraska, Lincoln, NE;Department of Computer Science and Engineering, University of Nebraska, Lincoln, NE;NASA Langley Research Center, Hampton, VA

  • Venue:
  • RV'10 Proceedings of the First international conference on Runtime verification
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Runtime verification has primarily been developed and evaluated as a means of enriching the software testing process. While many researchers have pointed to its potential applicability in online approaches to software fault tolerance, there has been a dearth of work exploring the details of how that might be accomplished. In this paper, we describe how a component-oriented approach to software health management exposes the connections between program execution, error detection, fault diagnosis, and recovery. We identify both research challenges and opportunities in exploiting those connections. Specifically, we describe how recent approaches to reducing the overhead of runtime monitoring aimed at error detection might be adapted to reduce the overhead and improve the effectiveness of fault diagnosis.