Improving Hit Ratio and Byte Hit Ratio Using Combined Pre-Fetching and Web Caching

Abstract

Combined caching and pre-fetching of web pages help to enhance bandwidth utilization, reduce access delay and also reduce the load on the server. Pre-fetching is the technique of fetching Web pages in advance that will be assumed to be needed or browsed in near future by the user and the caching process stores them in the memory. In the literature, number of works has been reported separately for caching and pre-fetching of Web pages. In this paper clustering based pre-fetching technique is combined with machine learning technique-SVM (Support Vector Machine)-LRU algorithm for Web proxy caching. With the help of real dataset it is demonstrated that the latter approach will be better than Clustering based pre-fetching technique with traditional caching policy like LRU in terms of bandwidth utilization and access latency.