Downloads: 117 | Views: 179

Review Papers | Computer Science & Engineering | India | Volume 4 Issue 8, August 2015

A Novel Algorithm Based on the Change in Entropy Value for Load Balancing in Cluster System

Mitali Mohire | Pratibha Yalagi ^[2] | Sulabha Apte

Abstract: The fast development in the internet era through the distributed network has changed the scope of the developers by improving task allocation and migration strategies. The problems in these systems are of ill-balanced task allocation strategies, long response time, low throughput rate and poor performance. The proposed load balancing algorithm focuses on a calculated entropy value of a load. To improve the performance of a system the load scheduling and migration policy are applied by considering the calculated entropy value. The load scheduling and migration is based on the factors like, speed of the system, server and client configuration, time of server-client communication, performance benchmark, etc. The proposed system focuses on a novel approach of calculating the entropy value of the load for the load balancing and migration. The traditional algorithm has the issues of low throughput, poor performance and long response time. The proposed algorithm achieves the load balancing and migration whenever required by calculating entropy and increases the system performance and throughput.

Keywords: Load balancing, change in value of entropy, cluster system

Edition: Volume 4 Issue 8, August 2015,

Pages: 986 - 989