MODELING AND ANALYSIS OF RELIABILITY PERFORMANCE FOR A DISTRIBUTED DUAL-PROCESSOR SYSTEM

This paper proposes an aggregate approach of extended stochastic Petri net and Markov renewal process to conduct behavior analysis and relia bility performance evaluation for distributed/parallel systems. Petri net is used because of its highly visual nature that can give insight into the nature of the modeled system. Markov renewal process is intro duced in order to increase the analytical power of Petri net model. Fu rther, numerical measure values are defined as indices to evaluate the reliability performance of the system. The modeling and analytical ap proach presented in this paper makes it possible to evaluate performan ce of a system with probabilistic behavior following non-exponential t ime distribution. The proposed approach could be effectively applied t o modeling and behavior analyzing for most asynchronous concurrent sys tems.