An Evaluation of Caching Strategies for Clustered Web Servers

Abstract

The exponential growth of the Internet and the increasing demands put upon Web servers has created the need for a scalable clustered Web servers. In this paper we attempt to analyse the important issues facing the implementor of a clustered Web server. The issues of TCP handoff problems, dynamically generated Web documents and how a Distributed Shared Memory paradigm might be used to optimise a clustered Web server. We analyse several clustered Web server architectures, and use traces from Web servers in a simulation of several of the Web server architectures to further examine the caching strategies of those architectures. Finally some recommendations are put forth for further research into areas that have been identified as critical to clustered Web servers.