Dynamic Event Driven Fault Diagnosis for Distributed Networks

Abstract

The occurrence of faults is a common feature in most networks and addressing this issue is an important aspect of network maintenance. In this project, we take up the problem of detecting faults in general arbitrary networks. The networks that we have considered for the implementation of our algorithm are general arbitrary networks; where every node in the network has a minimum of two neighbors. Various types of faults occur in such networks. This algorithm is applicable to hard faults. While doing so, we have assumed our own set of well-defined assumptions which will be followed during the implementation phase of the algorithm. The approach taken to diagnose the system brings about a system level diagnosis and consists of two phases. Each of phases has its own complexity and constraints. In this algorithm we follow the principle of periodic testing that occurs after a certain fixed period has passed. During each such period the tester node performs a test on a fault free node and then determines whether the node it has tested is faulty or fault free. This is a brief abstraction of the first phase also known as the ”Test Phase”.