用于超级节点网络的双缓存技术  

Double cache method for super-peer networks

作　　者：陈柏生[1] 杜吉祥[1] 

机构地区：[1]华侨大学计算机科学与技术学院,福建泉州362021

出　　处：《华中科技大学学报（自然科学版）》2011年第1期24-27,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60805021);福建省自然科学基金资助项目(A0810010)

摘　　要：针对P2P系统节点局部视图和用户查询不均衡的问题,提出一种用于超级节点网络的文件索引和数据双缓存技术.依据局部文件流行度统计,将用户感兴趣的资源索引缓存到本地超级节点,从而将单个节点视图扩展至全局;同时缓存拥有高访问率的远程热点文件,因此用户对热点文件的访问在本地即可获得响应.实验结果表明双缓存技术显著地提高了P2P系统的查询效率和数据传输速度,并可有效降低热点节点负载.The imbalance between local view of resources distribution and users′ requests was the reason why performance of peer to peer systems decrease.A fingers and files double cache method for super-peer networks was presented.Based on local file popularity statistics,the fingers of interests were dynamically cached to the native super peers;it consequently extended peers′ view of resources distribution globally.Files of high requests were also cached;local users could access the hot files from the native super peers.The experiment results show that the double cache method can dramatically improve the requests efficiency and download speed of the peer-to-peer system,and can effectively reduce the request load of hot spots.

关 键 词：对等系统 超级节点网络 负载均衡 文件流行度 缓存 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]