Toggle Main Menu Toggle Search

ePrints

Failure Detection in Replicated Systems

Lookup NU author(s): Emeritus Professor Santosh Shrivastava

Downloads


Abstract

Replicated processing with voting represents an attractive strategy for achieving reliability in real time systems. This technique permits N-Modular Redundant (NMR) nodes to be robust with respect to component failures. Three types of failures are identified in a replicated distributed system and the paper discusses how these failures can be detected as exceptions by majority voters. These exceptions include sequence exceptions caused by out of order message processing and hardware exceptions caused by processor and node failures. In particular, standard and exceptional domains for majority voters are defined and implementation strategies for the detection and handling of the discussed exceptions are presented.


Publication metadata

Author(s): Mancini LV, Shrivastava SK

Publication type: Report

Publication status: Published

Series Title: Computing Laboratory Technical Report Series

Year: 1987

Pages: 19

Print publication date: 01/06/1987

Source Publication Date: June 1987

Report Number: 238

Institution: Computing Laboratory, University of Newcastle upon Tyne

Place Published: Newcastle upon Tyne

URL: http://www.cs.ncl.ac.uk/publications/trs/papers/238.pdf


Share