• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2008, Vol. 30 ›› Issue (8): 22-25.

• 论文 • 上一篇    下一篇

HiTM:一种高速流量采集方法

王宏 龚正虎 殷泰晖   

  • 出版日期:2008-08-01 发布日期:2010-05-19

  • Online:2008-08-01 Published:2010-05-19

摘要:

针对目前高速流量采集存在的问题,对被采样的数据使用Hash算法压缩流ID长度,并在本文提出了一种高速流量采集方法HiTM。HiTM多级过滤器前增加一个高速缓存,减少流 ID处理时的访存次数和多级过滤器的报文处理数目,提高处理的效率。理论和实验表明,HiTM与目前的经典算法相比具有错误率低、访存次数少、统计速度快的特点,可以应  应用于高速网络流量采集。同时,算法的内存需求可以预先给定,在网络规模扩大时,算法的精度也可以得到很好的控制。

关键词: 流量采集 多级过滤器 流量测量

Abstract:

Based on analysing the characteristics of the Internet traffic, this paper gives a new multistage filter algorithm called HiTM to separate the traffic of large flows. The algorithm effectively reduces false positive errors and memory accesses. Theoretical analyses and experiments show that it can quic kly detect large flows using only few memories and memory accesses, In addition, this algorithm can be used in high-speed networks. The memory requireme nts can be predicted. The precision of the algorithm can also be well controlled when applied to large networks.

Key words: flow data collection;multi-stage filter;traffic measurement