针对网络用户行为的特点提出利用多尺度样本熵的两个评价规则来分析P-P flow流量矩阵,该方法能够对网络流量表现出来的一些特征进行合理解释,并得到一些有趣的结论。随后在多尺度样本熵分析特性基础上提出一种评估流量时间序列健康度的方法,通过对真实网络流量和仿真流量的实验,表明该方法简单有效,能够准确合理评估网络流量的健康程度。
For the features of network user behavior ,two evaluation rules of multi-scale sample entropy are used to analyze P-P flow traffic matrix .The method can reasonably explain some of the features shown in the network traffic and draw some interesting conclusions .Then a method to evaluate healthy degree of traffic time series is proposed based on analysis capacity of multi-scale sample entropy . Experiments conducted with actual network traffic and simulation network traffic show that the proposed method is simple and effective ,and can accurately evaluate the healthy degree of network traffic .