基于代理服务器的Web智能缓存系统研究与设计

发布时间：2018-05-13 18:06

本文选题：Web代理 + Cache系统　；参考：《北京邮电大学》2013年硕士论文

【摘要】：近几年来,互联网一直处于高速的发展状态,这种高速发展体现在多方面,如网民数量的急剧增多,网络提供的服务内容的多样化,以及网络中用户对网络应用种类和质量需求的提升等。这种高速的发展也带来了很多问题,在有限带宽的主干网络和有限资源服务器中由于各种各样的原因可能并不能够使得用户在发出请求后在满意的时间内得到反馈。因此在原本单一的网络服务模式中引入了中间节点的缓存系统服务模式。本文在原有Web端缓存系统的基础之上,对原本的代理缓存服务器缓存管理、替换、预取等策略进行深入研究和分析,由此引入了基于区域热点扩散的新缓存管理算法,同时在单点缓存代理中针对两种情况分别对缓存空间设计了缓存空间节约算法,实现了细粒度文件级的缓存节约策略。然后对目前缓存预取模型的更进一步研究,从基于区域“热点”发现的预取模式方面进行深入挖掘研究。最后,设计实现了新型的Web智能缓存系统。
[Abstract]:In recent years, the Internet has been in a state of rapid development, which is reflected in many aspects, such as the sharp increase in the number of Internet users and the diversification of the service content provided by the network. And users in the network of network application types and quality requirements, and so on. This kind of rapid development also brings many problems, in limited bandwidth backbone network and limited resource server, for various reasons, users may not be able to get feedback within a satisfactory time after making a request. Therefore, the caching system service mode of intermediate nodes is introduced into the original single network service mode. On the basis of the original Web cache system, this paper makes a deep research and analysis on the original proxy cache server cache management, replacement, prefetching and so on, and then introduces a new cache management algorithm based on the diffusion of regional hot spots. At the same time, a cache space saving algorithm is designed for two cases in the single point cache proxy, and the fine-grained file level cache saving strategy is implemented. Then, the current cache prefetching model is further studied, and the mining of prefetched patterns based on regional "hot spots" is carried out. Finally, a new Web intelligent cache system is designed and implemented.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP333

【参考文献】