Quick Consensus Through Early Disposal of Faulty Processes

摘要

This work reports an efficient solution for reaching agreement (consensus) among the processes of a distributed system. The better efficiency is achieved through early disposal of faulty processes while approaching for a consensus. The introduced network partitioning scheme further facilitates the progress by reducing the message exchange overhead. Simulation results establish that the proposed solution significantly reduces the message exchange complexity, in comparison to the schemes reported so far, simultaneously ensuring the fault-tolerance ability of a system as that of the known best results.