Research And Application Of Agent Based Cluster System Load Balance

With rapid development and widely usage of network technology, customer'srequirement in performance and security of applications has become more and morerigorous. The traditional way that upgrading hardware to improve application'sperformance can not match the requirement. Due to low cost, high reliability andthroughput and rich resource, cluster computing technology has become more andmore widely used. In order to enhance efficiency of load balance mechanism,communication andexpansibility of cluster system, At first, this paper discusses the present situation anddevelopment of cluster and Load Balance technology and brings forward theproblems and limitation of current cluster load balance mechanism, and also analysesthe characteristic of agent technology. Then, based on the synthetically analysis of theagent technology and cluster computing technology, an agent-based load balancemodel——ABLB is presented, which integrates the intelligent agent with cluster loadbalance mechanism. The model's targets, working principle, architecture, functionsand relationship of the model's each part are discussed subsequently, application ofABLB to the "Flight Optimizer and Profit Maximizer System"(FOPMAX) andexpatiates the project of FOPMAX's scheduler ABTS are also discussed. Additionallythis paper analyses and evaluation of the model and improvement to the ABLB aregiven.