A Site-Based Proxy Cache

In traditional proxy caches, any visited page from any Web server is cached independently, ignoring connections between pages. And users still have to frequently visit indexing pages just for reaching useful informative ones, which causes significant waste of caching space and unnecessary Web traffic. In order to solve the above problem, this paper introduced a site graph model to describe WWW and a site-based replacement strategy has been built based on it. The concept of "access frequency" is developed for evaluating whether a Web page is worth being kept in caching space. On the basis of user's access history, auxiliary navigation information is provided to help him reach target pages more quickly. Performance test results have shown that the proposed proxy cache system can get higher hit ratio than traditional ones and can reduce user's access latency effectively.