Web Prefetching And Caching The Integrated Model

Presently caching and prefetching techniques are the primary solutions used to reduce Web access latency. Web caching technique has been widely used in different places of Internet. But the cache mechanism only uses the time locality of WWW access pattern, it can't cache the documents which are not accessed before, so the quality of response can't be improved much yet. As a complement way, Web prefetching is the most effective method to break the upper bound of caching performance. Prefetching is one kind of active cache that can cache the pages which are still not requested by the user, which is an expansion from the time locality to the space locality. There are two problems to be solved before prefetchingcan be put into practice: One is prediction——speculate which Web objects will be mostlikely requested before the users take the action. The other is prefetching control. Its effect will directly determine whether prefetching can be used widely and will not have bad effect on the network performance. So prefetching should determine which Web objects will be prefetched and how many Web objects can be prefetched according to the system situation. Web caching and prefetching which are integrated will highly reduce the latency in Web applications. This thesis analyzes the prefetching control policy based on prefetch-cache interaction, and also presented prefetchability concept and measurement method of Web object.First, this thesis gives the problems Internet faced and corresponding solutions; and describes the concept, classification and structure of Web caching and prefetching; then summarizes existing cache replacement algorithms, prediction algorithms and prefetching control strategies.Then relative knowledge about sequential pattern is introduced and the mining mechanism that adopts depth-first search strategy for prediction model is presented. Compared with previous breadth-first search strategy, this mining mechanism that adopts depth-first search strategy base on lexicographic tree and uses the data structure of bitmap to save and calculate the support of sequences fast.Then prefetchability concept and measurement method of Web object are presented. And further, three prefetch model based on the prefetch-cache interaction are established. The cost function of prefetching is analyzed to conclude the prefetch threshold. By prefetching these items that have access probability higher than the threshold, we can find the tradeoff between the reduced latency and the potential increase in network and resources loads. Simulations show that the prefetch-cache interaction based prefetching control model can manage prefetching effectively, and can reduce the excessive consumption of the system...