大型文件共享系统的用户行为分析  

Analysis of User Behaviors for Large Scale File Sharing Systems

作　　者：张成彬[1] 涂旭平[2] 

机构地区：[1]江苏盐城工学院信息工程学院,盐城224051 [2]华中科技大学计算机学院,武汉430074

出　　处：《武汉理工大学学报》2009年第18期32-34,共3页Journal of Wuhan University of Technology

摘　　要：大型P2P文件系统是当前互联网的一种杀手级应用之一,其吞噬了大量的网络带宽。本文从用户行为包括用户下载次数、文件下载次数、文件类型、文件大小对此类型的P2P应用进行了分析。分析表明,用户下载次数、文件下载次数的分布呈现Zipf分布;音乐文件类型被下载次数最多,但是媒体文件却下载的总流量最大,其中rmvb文件最为流行;80%的文件大小在500 M以内。Large scale P2P-based file sharing system is one kind of the killer-applications of the Internet nowadays because it consumes too much bandwidth. This paper analyzes the user behavior pattern of this kind of systems from the trace of a famous deployed system. The result shows that, user downloading times and file downloading times have a Zipf distribution. The music flies are most frequently downloaded but the media file dominates network traffic. File type RMVB is the most popular media file type. For file size distribution, 80% fries are less than 500M.

关 键 词：文件共享 P2P 用户行为 大型文件共享系统 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]