基于对象替换与预取的Web缓存模型研究

发布时间：2018-05-31 11:13

本文选题：Web缓存 + 协同过滤　；参考：《湖南科技大学》2014年硕士论文

【摘要】：随着Internet网络技术的迅猛发展，网络接入用户量和信息数据量急剧增加，互联网已经深入渗透至社会生活的各个方面，深刻地影响着人类的生产方式与生活习惯。由于不同网络拓扑结构具有明显的差异、多媒体内容在网络资源中的比重不断增加以及网络带宽的限制，，这些因素导致网络传输具有较高延时，影响了用户网络访问体验。针对上述问题，研究人员提出了Web缓存技术与Web预取技术，这两种技术都能极大地提高网络带宽利用率、减轻网络出入口吞吐容量并减少用户访问延时，从而改善用户网络服务体验。本文研究了Web缓存技术与Web预取技术，并将两者进行有机结合，提出了一种新的Web缓存模型，该模型的缓存模块为本文提出的改进型缓存替换算法，预取模块采用了基于用户访问模式的预取算法。论文主要完成了如下几个方面的具体研究：首先，通过研究和分析已有的Web缓存替换算法，设计了一种基于协同过滤的Web缓存替换算法GDSF-CF（Greedy Dual Size Frequency Collaborative Filtering）。该算法利用协同过滤技术生成Web对象的预测访问频率，并考虑了用户的访问特征与Web对象的属性特征，通过目标函数的计算确定Web对象的缓存价值；当存储空间需要进行替换操作以容纳新的Web对象时，则对每个Web对象的缓存价值进行比较，替换出价值最小的对象，从而完成替换过程，维护代理服务器的可用性。然后，提出了一种基于对象替换与预取的Web缓存模型MGP（Model of GDSF-CFwith Prefetching）。该模型包含缓存与预取两大模块，缓存模块为本文提出的GDSF-CF替换算法，预取模块采用了基于用户访问路径分析的预取算法，该预取算法通过分析用户的访问序列形成目标网页的预取集合；当用户请求到达代理服务器时，缓存模块与预取模块将共同处理用户请求，有效兼顾了用户访问的时间局部性与空间局部性。最后，通过仿真实验对本文提出的GDSF-CF算法与MGP模型进行了评价与分析，验证了GDSF-CF算法与MGP模型的有效性。
[Abstract]:With the rapid development of Internet network technology, the number of network access users and the amount of information have increased dramatically. The Internet has deeply penetrated into all aspects of social life, and has deeply affected the mode of production and living habits of human beings. Due to the obvious differences in different network topologies, the increasing proportion of multimedia content in network resources and the limitation of network bandwidth, these factors lead to high delay of network transmission and affect the user's network access experience. In view of the above problems, researchers put forward Web cache technology and Web prefetching technology, both of which can greatly improve the utilization of network bandwidth, reduce the throughput capacity of network entry and exit, and reduce user access delay. In order to improve the user network service experience. In this paper, Web cache technology and Web prefetching technology are studied, and a new Web cache model is proposed. The cache module of this model is an improved cache replacement algorithm proposed in this paper. Prefetching module adopts prefetching algorithm based on user access mode. The thesis mainly completed the following aspects of the specific research: Firstly, by studying and analyzing the existing Web cache replacement algorithms, a Web cache replacement algorithm based on collaborative filtering, GDSF-CF(Greedy Dual Size Frequency Collaborative filtering algorithm, is designed. The algorithm uses collaborative filtering technology to generate the predicted access frequency of Web objects, and considers the access characteristics of users and the attribute features of Web objects, and determines the cache value of Web objects through the calculation of objective functions. When the storage space needs to be replaced to accommodate the new Web object, the cache value of each Web object is compared to replace the least valuable object, thus the replacement process is completed and the availability of the proxy server is maintained. Then, a MGP(Model of GDSF-CFwith prefetching model based on object replacement and prefetching is proposed. The model includes two modules: cache and prefetching. The cache module is the GDSF-CF replacement algorithm proposed in this paper. The prefetching module adopts the prefetching algorithm based on user access path analysis. The prefetching algorithm forms a prefetching set of the target web pages by analyzing the user's access sequence. When the user requests to reach the proxy server, the cache module and the prefetching module will jointly process the user requests. The time locality and spatial locality of user access are taken into account effectively. Finally, the GDSF-CF algorithm and MGP model proposed in this paper are evaluated and analyzed by simulation experiments, and the validity of GDSF-CF algorithm and MGP model is verified.
【学位授予单位】：湖南科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP333

【参考文献】