一种基于小波谱的流量识别方法

摘要

针对流量识别方法受到服务端口、协议签名、流量加密等限制的问题,提出一种基于多尺度小波谱的流量识别方法.利用离散小波变换对原始网络流量进行多尺度分解,分别计算不同尺度下的小波谱,使用K-means算法对这些小波谱进行聚类分析,由此实现网络流量的识别.实验结果表明,该方法具有较高的识别准确率.%Aiming at the problem of current traffic identification method limited by service port, protocol signatures and traffic encryption, this paper proposes a new traffic identification method based on multi-scale wavelet spectrum. The new method is able to achieve the identification of different network traffic by performing clustering analysis of K-means algorithm over the different scales wavelet spectrum coefficients that are inferred from the raw traffic by using Discrete Wavelet Transform(DWT). Experimental result shows that it has accurate identification results.