Toggle Main Menu Toggle Search

Open Access padlockePrints

An Empirical Study of the Performance of Distributed Replicated Systems

Lookup NU author(s): Dr Paul EzhilchelvanORCiD, Emeritus Professor Isi Mitrani, Emeritus Professor Santosh Shrivastava

Downloads


Abstract

Replicated processing with majority voting - N modular redundant processing - is a well known method of achieving fault tolerance. We consider a distributed system where a job can be broken into a number of sub-jobs which are processed sequentially at various processors of the distributed system. The performance of such a system is then compared with the replicated version of the system where each sub-job will require concurrent replicated processing with majority voting. The objective is to examine the relative impact of voting times and processor failure probabilities on the overall performance. A variety of system configurations are evaluated by means of computer simulations, and also by analytical approximations. The accuracy of the latter is examined.


Publication metadata

Author(s): Ezhilchelvan PD, Mitrani I, Shrivastava SK

Publication type: Report

Publication status: Published

Series Title: Computing Laboratory Technical Report Series

Year: 1989

Pages: 25

Print publication date: 01/01/1989

Source Publication Date: 1989

Report Number: 278

Institution: Computing Laboratory, University of Newcastle upon Tyne

Place Published: Newcastle upon Tyne

URL: http://www.cs.ncl.ac.uk/publications/trs/papers/278.pdf


Share