A Comparison of Cluster and Grid Configurations Executing Image Processing Tasks in a Local Network

Abstract

The use of clusters and grids for high-performance applications has become widespread lately. In general, one associates clusters with dedicated LANs and grids with WANs and internets. This paper reports the results of a comparative analysis between cluster and grid performance when executing a load-balanced, image processing application in a local-area network. The results obtained show almost no difference in total execution times, indicating that grid software may be used advantageously in LANs to run similar load-balanced applications.