Research And Application On Integration Of Web Caching And Prefeching

With the develpoment of Web2.0, Micro-blog has been becoming the most popular application on the Internet. The sina microblog already has hundred million registered users, the daily amount of micro-Blog up to tens of millions. Such a large amount of information will inevitably lead to greater visits and be followed by the load is increased at all levels of server and network access delay. These issues will unconditional reduce the satisfaction user and impact on the further development of micro-blog. The key is to solve the above problem of Web data caching and prefetching techniques. Cache can effectively reduce the time and distance on data acquisition by store data in the nodes of network. Prefetching can effectively reduce the waiting time of users’ access by predicting and prefetching the future data which would be accessed. However, the current study for these two technologies is not good enough to deal with this new data features caused by micro-blog. Micro-blog has a strong time characteristics, and each micro-blog has a fixed size.Therefore, this paper proposed a cache replacement algorithm based on time constraints and data prefetching and improved data prefetching algorithm, meanwhile integrate the two algorithms effectively to improve the access efficiency and reduce access time.In this paper, we proposed a cache replacement algorithms based on comprehensive consideration of the data object access time, size, and prediction by researching the traditional cache replacement algorithm. The cache replacement algorithm will determine what data will be the replacement off the cache space. This algorithm fully considered the feature of micro-blog, and replace the data that has large size and been accessed for a long time to keep data "fresh". At the same we improved the prefixSpan sequential pattern mining algorithm by adding Time constraints. Then we combined the frequent sequence with the current accessing sequence, and predict user access behavior through sliding window method. Final we proposed the integration of Web caching and prefetching model in oreder to combine these technologies which will achieve better improvement results.In the last part of this paper simulate the integration of Web caching and prefetching model by useing the real log data. By comparing with the efficiency of several other cache replacement algorithms, we succeed to verify good performance of the proposed integration of caching and prefetching model.