PORTABLE CONTROLLING SOFTWARE PACKAGE FOR PARALLEL FAULT-TOLERANT COMPUTER SYSTEM

摘要

A portable software system to support distributed parallel and fault tolerant computing in a PC network environment is described.It facilitates better utilization of system resources and provides a platfom for fault tolerant computing.The software system consists of a concurrent control library and a program translator.The design is dedicated for a master-slave.loosely coupled and Tandemmodel configuration.The implementation of slave system management,communication protocol and fault tolerance is discussed.Some performance analysis using networked PC's is also addressed.