期刊名称：International Journal of Engineering and Computer Science
印刷版ISSN：2319-7242
出版年度：2014
卷号：3
期号：10
页码：8752-8764
出版社：IJECS
摘要：This work is an extension of our previous work on task scheduling of a Distributed Computing Software inthe presence of faults [2] in which an attempt was made to identify the task scheduling algorithms used fordistributed environments that perform well in the presence of faults due to network failure or processorfailure in the distributed system. In this paper we give some extensive results for identifying the taskscheduling algorithms that perform well in the presence of communication and computation faults present insome special task graphs like systolic array task graphs, Gaussian elimination task graphs, Fast Fouriertransform task graphs, and divide-and-conquer task graphs which can represent the distributed software. Forour study we have selected six task scheduling algorithms. We have compared these algorithms using threecomparison parameters like normalized schedule length, number of processors used and average runningtime, and evaluated them on the above mentioned task graphs in the presence of communication andcomputation faults. It is further evaluated under random and constant fault delays.
