Identification of Crash Fault & Value Fault for Random Network in Dynamic Environment

Abstract

During the past few years distributed systems have been the focus of considerable research in computer science. Fault tolerance in distributed systems is a wide area with a significant body of literature that is vastly diverse in methodology and terminology. Fault tolerance is the ability of a system to perform its function correctly even in the presence of internal faults. An extensive methodology has been developed in this field over the past few years, and a number of fault-tolerant machines have been developed but most dealing with random hardware faults, while a smaller number deal with software, design and operator faults to varying degrees.
Our work mainly focuses on the simulation of the system that deals with software faults means the faults that occur because of the failure or error in the internal software component. Our work is restricted to distributed diagnosis in dynamic fault environment. Basically we have created different not-completely connected random networks with number of nodes ranging from 8 to 256.Then we have induced faults to these networks dynamically using poison distribution. Three different algorithms have been implemented to detect the faults and the comparison among these algorithms, based on delay latency and number of message exchanges, has been represented graphically. The software faults that we had dealt with are crash fault and value fault in a distributed system (not-completely connected network). Although many researches have been done in the crash fault area but very less work has been done in diagnosing the value faults in dynamic fault environment.