sFlow网络流量分析算法研究  

Research on Algorithm of the sFlow Network Traffic Analysis

作　　者：宋松滋[1] 林南晖[2] 何俊[1] 

机构地区：[1]华南师范大学计算机学院,广州510631 [2]华南师范大学网络中心,广州510631

出　　处：《现代计算机》2009年第4期86-87,96,共3页Modern Computer

摘　　要：sFlow是一种"统计采样技术",其分析方法采用1/n数据包采样来描述网络流量的特征,随着获取的样点数量的增加,相对采样误差减小,最终将趋向能与收集每一个数据包来描述网络流量特征达到同样的结果。通过详细地研究sFlow技术,在对比的情况下总结了其优缺点,并对sFlow技术的发展方向提出展望。sFlow is a kind of ＂statistical sampling technique＂, which describes the characteristics of network traffic by analyzing 1/n packet samplles. With the increasing in the number of the samples obtained, the relative sampling error decreases, which will eventually tend to achieve the same results as describing the characteristics of network trafficwork by collecting every single data packet. Gives a full description of the popular sFlow technique as well as its pros and cons, proposes a blueprint of its further development.

关 键 词：SFLOW技术 网络流量监测 流量分析 

分 类 号：TP393.06[自动化与计算机技术—计算机应用技术]