Web prefetch performance evaluation in a real environment

Abstract

Web prefetch is a technique aimed at reducing the user's perceived latency by downloading, during the navigation idle times, the web objects before the user asks for them. Despite different research efforts can be found in the literature about this subject there are few implementations for real environments. We proposed a real framework that provides web prefetching on current web client and server software working under the standard HTTP 1.1 protocol. This framework also provides detailed statistics which are very convenient for performance evaluation studies. In this paper we evaluate and compare the performance of different prediction algorithms in real conditions showing the usefulness of our proposal for this kind of environments.