Content-Based Distribution for Load Sharing in Locally Clustered Web Servers

Abstract

A cluster consists of a collection of interconnected stand-alone computers working together and provides a high-availability solution in application areas such as web services or information systems. With the growing popularity of the Internet, services using the World Wide Web are increasing, and it is certain that web server clusters will be the basic architecture. The Layer-4 distribution algorithms of web server clusters have widely studied, but they are content information blind. They select the target server when the client establishes the TCP/IP connection. On the contrary, content-based load distribution uses the detailed data found in the application layer to intelligently route user requests among web servers. In this paper, we propose an effective content-based load distribution algorithm that considers cache hit and load information of the web servers under the web server clusters. In addition, we expand this algorithm in order to manage user requests for dynamic file. Specially, our algorithm does not keep track of any frequency of access information or try to model the contents of the caches of the web servers.