Journals Proceedings

Abstract

This paper addresses solution to the problem of reaching an agreement (consensus) in a distributed system. The aim is to maximise fault tolerance as well as reduce the message exchange overhead. The proposed algorithm uses as few as two rounds of message exchange. This better efficiency is achieved through early disposal of faulty processes. The network partitioning scheme proposed later further reduces message exchange overhead and thus network traffic Keywords- Consensus, Byzantine Agreement, Reaching Agreement, Early Stopping, Decision Vector