Web缓存与预取系统的设计与实现

发布时间：2018-10-26 09:17

【摘要】：在Internet飞速发展的今天，随着网络信息和用户接入数量的快速增长，已有的网络带宽已经不能够很好地满足用户的服务请求，怎样提升用户请求访问的速度和网络的整体性能已经变成了一个急需解决的问题。缓存和预取技术是一个有效的解决方法，基于此针对Web的访问特性和用户的感兴趣度对缓存置换算法进行了改进，并将Web缓存和预取技术相结合，有效提高了缓存系统的命中率，改善了网络性能。现有的Web缓存置换策略大都只是考虑了获取Web对象本身的代价，参考因素比较单一，，导致缓存命中率较低。论文首先分析了用户Web访问行为的特性，并结合用户的感兴趣程度和Web对象所属的类型，在GDSF算法的基础上，增加了对这几种因素的考虑，提出了一种新的Web缓存置换策略GDSFI，并通过仿真实验对其进行了性能验证和比较。 Web预取是对缓存技术的一种有效补充，通过对用户行为的预测能够在浏览时间内完成下一Web对象的请求。在设计预测算法时，提取出用户访问最热门的N个网站，通过对访问日志进行统计得出热门领域中最受欢迎的Web列表，作为预取的依据。将预取和缓存相互结合，在缓存某一Web对象的同时判断是否需要进行预取操作，从而可以实现更高的缓存性能。通过对缓存置换策略的改进和缓存预取操作相结合，能够有效提高系统的缓存命中率和字节命中率，改善网络的性能和服务质量。
[Abstract]:With the rapid development of Internet, with the rapid growth of network information and the number of users' access, the existing network bandwidth can no longer meet the service requests of users. How to improve the speed of user request access and the overall performance of the network has become an urgent problem. Caching and prefetching technology is an effective solution. Based on the access characteristics of Web and the interest of users, the cache replacement algorithm is improved, and the Web cache and prefetching technology are combined. The hit rate of cache system is improved effectively and the network performance is improved. Most of the existing Web cache replacement strategies only consider the cost of obtaining Web objects, and the reference factor is single, which leads to low cache hit rate. This paper first analyzes the characteristics of user's Web access behavior, and combines the degree of interest of user and the type of Web object, on the basis of GDSF algorithm, adds consideration to these factors, and proposes a new Web cache replacement policy GDSFI,. The performance is verified and compared by simulation experiment. Web prefetching is an effective supplement to cache technology. By predicting user behavior, the request of the next Web object can be completed within browsing time. In the design of prediction algorithm, users are extracted from the most popular N websites, and the most popular Web list in the popular field is obtained by statistics of the access log, which is the basis of prefetching. By combining prefetching and caching, we can cache a Web object and decide whether we need to prefetch or not, so that we can achieve higher cache performance. Through the combination of cache replacement strategy and cache prefetch operation, the cache hit rate and byte hit rate of the system can be improved effectively, and the network performance and quality of service can be improved.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.07

【参考文献】