An experimental study of prefetching and caching algorithms for the World Wide Web

Caching and prefetching have often, been studied as separate tools for enhancing the access to the World Wide, Web. The goal of this work is to propose integrated Caching and Prefetching Algorithms for improving the performances of web navigation. We propose a new prefetching algorithm that uses a limited form of user cooperation to establish which documents to prefetch in the local cache at the client side. We show that our prefetching technique is-highly beneficial only if integrated with a suitable caching algorithm. We consider two caching algorithms, Greedy-Dual-Size [6,17] and Least Recently Used, and demonstrate on trace driven simulation that Greedy-Dual-Size with prefetching outperforms both LRU with prefetching and a set of other popular caching algorithms.