Distributed Load Balancer Using Estimation Algorithm

Abstract

The growth of distributing systems with the possibility of sharing the available resources has led to the ability of executing some jobs arrived at a certain node remotely. Consequently, load balancing is aiming at reducing the overall response time of jobs execution in any distributed system. Load Balancing is the process of improving the performance of parallel and distributed system. This system presents the distributed load balancing system by estimation algorithm. In this system, distributed model is designed with clusters which have processing nodes. Processing nodes in each cluster have computation processes for processing tasks and each cluster has its own load balancer. The load balancers use estimation algorithm to get the workload of the clusters. Based on the estimated workload, the decision making process decides tasks should be transferred or not. Then, the load balancer estimates how much tasks to be transferred and perform the task transferring process.