摘要
Web代理服务器缓存能够在一定程度上解决用户访问延迟和网络拥塞问题,Web代理缓存的缓存替换策略直接影响缓存的命中率,从而影响网络请求响应的效果;为此,使用一种通过固定大小的循环滑动窗口提取Web日志数据的多项特征,并使用高斯混合模型对Web日志数据进行聚类分析,预测在窗口时间内可能再次访问到Web对象,结合最近最少使用(LRU)算法,提出一种新的基于高斯混合模型的Web代理服务器缓存替换策略;实验结果表明,与传统的缓存替换策略LRU、LFU、FIFO、GDSF相比,该策略有效提高了Web代理缓存的请求命中率和字节命中率。
Web proxy cache can solve the problems of user access delay and network congestion to a certain extent.The cache replacement strategy of web proxy cache directly affects the hit rate of cache,thereby affecting the effect of network request response.To solve this problem,using a fixed-size sliding window to extract multiple features of Web log data,and using a Gaussian mixture model for cluster analysis of Web log data,predicting that the Web object may be accessed again within the sliding window time,combining the least using(LRU)algorithm,a new cache replacement strategy of web proxy server based on Gaussian mixture model is proposed.The results show that compared with the traditional cache replacement strategies such as LRU,LFU,FIFO,GDSF,the proposed strategy effectively improves the request hit rate and byte hit rate of web proxy cache.
作者
唐榜
吴珏
杨福军
杨雷
Tang Bang;Wu Jue;Yang Fujun;Yang Lei(School of Computer Science and Technology,Southwest University of Science and Technology,Mianyang 621000,China;Institute of Computational Aerodynamics,China Aerodynamics Research and Development Center,Mianyang 621000,China)
出处
《计算机测量与控制》
2021年第2期166-170,175,共6页
Computer Measurement &Control
基金
国家数值风洞工程支持项目
国家重点基础研究发展计划基金项目(2014CB744100)
西南科技大学博士基金(13zx7102)。
关键词
WEB缓存
替换策略
循环滑动窗口
高斯混合模型
访问预测机制
Web cache
replacement strategy
circular sliding window
Gaussian mixture model
access prediction mechanism
