Overview of the Architecture

Abstract

Despite the enormous improvements in the quality of computer technology that have been made during the last few decades, failures of computer system components cannot be ruled out completely. Some application areas therefore need computer systems that continue to provide their specified service despite component failures — such computer systems are said to be fault-tolerant. Distributed computing systems — in which multiple computers interact by means of an underlying communication network to achieve some common purpose — offer attractive opportunities for providing fault-tolerance since their multiplicity of interconnected resources can be exploited to implement the redundancy that is necessary to survive failures.