Fault Detection and Diagnosis in Distributed Systems: An Approach by Partially Stochastic Petri Nets

  • Authors:
  • Armen Aghasaryan;Eric Fabre;Albert Benveniste;René/e Boubour;Claude Jard

  • Affiliations:
  • IRISA/INRIA, projet Sigma 2, Campus de Beaulieu, F-35042 Rennes cedex, France/ aaghasar@irisa.fr;IRISA/INRIA, projet Sigma 2, Campus de Beaulieu, F-35042 Rennes cedex, France/ fabre@irisa.fr;IRISA/INRIA, projet Sigma 2, Campus de Beaulieu, F-35042 Rennes cedex, France/ benvenis@irisa.fr;France TELEcom/CNET Lannion - DTL/DLI, Technopole Anticipa, 2, av. Pierre Marzin, F-22307 Lannion cedex, France/ renee.boubour@cnet.francetelecom.fr;IRISA/CNRS, projet Pampa, Campus de Beaulieu, F-35042 Rennes cedex, France/ jard@irisa.fr

  • Venue:
  • Discrete Event Dynamic Systems
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

We address the problem of alarm correlation in large distributedsystems. The key idea is to make use of the concurrence of events in orderto separate and simplify the state estimation in a faulty system. Petri netsand their causality semantics are used to model concurrency. Specialpartially stochastic Petri nets are developed, that establish some kind ofequivalence between concurrence and independence. The diagnosis problem isdefined as the computation of the most likely history of the net given asequence of observed alarms. Solutions are provided in four contexts, with agradual complexity on the structure of observations.