Comparison of load balancing algorithms for clustered web servers

Abstract

The rapid growth of the World Wide Web has brought huge increase in traffic to popular web sites. As a consequence, end users often experience poor response time or denial of service. A cluster of multiple servers that behaves like a single host can be used to improve the throughput and alleviate the server bottlenecks. To achieve such a cluster, we need robust routing algorithms that provide scalability, effective load balancing and high availability in a constantly changing environment. Due to an ever-increasing diversity of workloads and cluster configurations, it is very difficult to propose a single algorithm that performs best under all conditions. In this paper, we review a number of proposed load balancing algorithms for clusters of web servers. We focus on an experimental analysis of the performance under a number of well-known load balancing algorithms. We study the performance of the algorithms through a simulation to evaluate their performance under different conditions and workloads. The results of our study are reported in the paper.