Discrete event approach to network fault management

Abstract

<p>Failure diagnosis in large and complex systems such as a communication network is a critical task. An important aspect of network management is fault management, i.e.,determining, locating, isolation, and correcting faults in the network. In the realm of discrete event systems Sampath et al proposed a failure diagnosis approach, and Jiang et al proposed an efficient algorithm for testing diagnosability. In this work, we adopt the framework of the communicating finite state machine (CFSM) of Miller et al for modeling networks and to investigate fault detection, fault identification and fault location using Sampath et al and Jiang et al methods. Our approach provides a systematic way of performing fault diagnosis aspects of network fault management.</p>