校园网网络流量自相似性的测定

Verification of the Self-Similarity of Campus Network Traffic

根据自相似性理论,网络流量自相似性的程度可由计算Hurst参数的大小来测定。本文通过在校园网主干路上进行网络数据包的采集,然后用方差时间方法和R/S方法计算所采集流量的Hurst参数值,以检测校园网网络流量的自相似性。通过实验的分析与比较,验证了自相似性的存在,并对如何运用Hurst值的变化来进行流量异常检测做了初步的探讨。

I he selt-simtlarity degree ot network trattic is checked by computing the Hurst parameter based on the selfsimilar theory. In this paper, we capture packets on the campus network＇s backbone, and then compute the Hurst parameter to check the self-similar degree of network traffic by using the R/S method and the variance-time method. The existence of self-similarity is verified by experiments. We also discuss the network traffic abnormal detection based on the Hurst parameter＇s obvious changes.