Towards software health management with bayesian networks

  • Authors:
  • Johann Schumann;Ole J. Mengshoel;Ashok N. Srivastava;Adnan Darwiche

  • Affiliations:
  • SGT Inc, NASA Ames, Moffett Field, CA, USA;Carnegie Mellon University, Moffett Field, CA, USA;NASA Ames, Moffett Field, CA, USA;University of California at Los Angeles, Los Angeles, CA, USA

  • Venue:
  • Proceedings of the FSE/SDP workshop on Future of software engineering research
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

More and more systems such as aircraft, machinery, and cars rely heavily on software, which performs safety-critical operations. Assuring software safety though traditional V&V has become a tremendous, if not impossible task, given the growing size and complexity of the software. We propose that SWHM (SoftWare Health Management) has the potential to increase safety and reliability of high-assurance software systems. SWHM can build upon the advanced techniques from the area of system health management to continuously monitor the behavior of software during operation, quickly detect anomalies and perform automatic and reliable root-cause analysis. Such a system would not replace traditional V&V, but rather supplement it. The information provided by the SWHM system can be used for automatic mitigation mechanisms (e.g., recovery, dynamic reconfiguration) or presented to a human operator for further analysis. SWHM may also feature a key prognostic capability, which can improve the reliability and availability of the software system because it provides information about soon-to-occur failures or looming performance bottlenecks. In this paper, we discuss research challenges associated with developing an SWHM system, and discuss how Bayesian networks (BN), a key technology used in advanced diagnostics systems may be used for SWHM modeling.