Intelligent prefetch in WWW using client behavior characterization

Improving the latency perceived by a web client has drawn a lot of attention right from the genesis of the world wide web. Hierarchical caching is found to improve the latency in addition to reducing the web traffic in the Internet backbone. In the recent times, the Uniform Resource Locators (URL) are found to be dynamic i.e., the contents of the URLs keep changing every few minutes or during every visit. This proliferation of dynamic URLs renders the static caching inefficient. In this paper, we present a prefetching technique by characterizing the web client browsing behavior and show improvement in the web client cache hit ratio in the presence of dynamic URLs. The features we considered are the trend in visit counts and the order of visits of URLs by the web client. The proposed technique shows an average of 13% client cache hit ratio even when all the visited URLs are dynamic.