Distributed Chronicles to Faults Recognition

Abstract

The chronicles paradigm has been used to determine fault in dynamic systems, allow modeling the temporal relationships between observable events and enabling to describe the patterns of behavior of the system. The mechanisms used until now usually use semi-centralized approaches, which consist of a central component that is responsible for making the final inference about the fault diagnosis of the system, based on the information collected from local diagnosers. Hence this model is not suited for monitoring very large systems. We propose in this article a fully distributed approach. This distributed chronicle recognition is illustrated in the context of e-commerce with a Service Oriented Application implementation