| The rapid development of the Internet today, with the explosion of information resources, making network information growth rate far exceeds the disk capacity and network speeds increaseing. In order to solve these problems, people made a variety of technical solutions, the chief among them is the Web Caching and Prefetching. At present, for prefetching and caching technology research only emphasized on the prefetching algorithm or replacement policy improvements, lack of the research of the combination of prefetch algorithms and a replacement policy.This article focuses on the following aspects from the indepth study and discussion.Firstly, in forecasting models, the sequential pattern mining applied to Web logs, making simple improvements on the PrefixSpan of sequential pattern algorithm. The algorithm is applied to sequence prediction based on user access model, using WAP-tree prefix tree store the user's access sequence, using the improved PrefixSpan algorithm for mining frequent sequential pattern, that is, the user's access characteristics. Combination the user's current access sequence with the predicted windows to form the forecast queue which will be accessed by users.Secondly, in the hit rate of replacement strategy, Based on the replacement algorithm GDSF, join the forecasting mechanism, proposed a cache replacement algorithm PGDSF, according to the user's current request for a visit and user access records of past history predict the object which will be accessed.Increase the hit rate of cache in certain buffer space.Again, based on the replacement algorithm PGDSF, presents a Web cache and prefetch the integration model, combining cache system with prefetch system, to improve the network quality of service.Finally, through simulation experiments to verify replacement algorithm PGDSF, show that the this cache replacement algorithm PGDSF in a certain space, compared to other replacement algorithm has a high document hit ratio and byte hit ratio. |
PDF Full Text Request