| Since its introduction, the World Wide Web has seen an exponential growth in its usage. The result is that a vast amount of information is available to a user at the click of a button. The growth in the popularity of the web, however, has produced its own problems in terms of annoying and frustrating delays in response. Reducing this delay has opened up a new and interesting area of research. The major contribution to this delay comes from two entities: network, and the web server. In this thesis, we focus our attention on the web server delay.; We present a technique that reduces the load imposed on web servers by increasing the hit ratio in server cache and thus reducing the client waiting time. This technique addresses important stages in effective cache management such as "which files to be cached?", "which files to be validated?", and "which files to be replaced?". This results in a more intelligent and productive utilization of server cache that, in turn, reduces the client waiting time. Three different cache replacement policies and two different new weight metrics are tested to obtain the best performance. Weight metrics are formulated using important factors of object history and objects in the cache are ranked by the weight. Results of experiments demonstrate that the proposed technique performs much better than existing techniques. The proposed algorithm represents an online, dynamic and adaptive technique, which provides improved hit ratio, latency and server utilization. |
