LDMA: Load balancing using decentralized decision making mobile agents

Abstract

This paper introduces a new load balancing algorithm, called LDMA (Load balancing using Decentralized decision making Mobile Agents), which distributes load among clustered web servers connected in a mesh topology, by a communications network and compares its performance with other load balancing algorithm: MALD (Mobile Agent based LoaD balancing). Architecture is developed for the same and all necessary attributes such as load deviation, system throughput and response time incurred as a result of the work are dealt with. In past works, a centralized decision making algorithm was used for dispatching requests to web servers in the distributed client/server environment. In the proposed approach, a decentralized decision making algorithm is used for distributing requests among web servers. The simulator is developed in C++ and its performance is evaluated. The analysis shows that LDMA is better than centralized decision making load balancing algorithms. © Springer-Verlag Berlin Heidelberg 2006.