THE STUDY OF LOAD BALANCING STRATEGIES FOR THE DISTRIBUTED COMPUTING SYSTEM

Abstract

Processing speed of a system is always highly intended. From the beginning of the development newlineof computer it is always focused on the system performance that is how to improve the speed or newlinethe performance of an existing system. Especially the business organizations, defense sectors and newlinescience groups need the high performance systems for constant change of their day to day need. newlineParallel and distributed computing system provides high performance environment that are able newlineto provide huge processing power. Distributed network is mainly heterogeneous in nature in the newlinesense that the processing power of nodes, network topology, communication medium, operating newlinesystem etc. may be different in different network which are even widely distributed over the newlineglobe. In order to get the maximum efficiency of a system the overall work load has to be newlinedistributed among the nodes over the network. The distribution of loads to the processing newlineelements is simply called the load balancing problem. In a system with multiple nodes there is a newlinevery high chance that some nodes will be idle while the other will be over loaded. The goal of newlinethe load balancing algorithms is to maintain the load to each processing element such that all the newlineprocessing elements become neither overloaded nor idle that means each processing element newlineideally has equal load at any moment of time during execution to obtain the maximum newlineperformance (minimum execution time) of the system. So the proper design of a load balancing newlinealgorithm may significantly improve the performance of the system...... newline