微信流量分类模型及其业务识别算法研究

摘要

Taking the WeChat traffic model and business identification as the main research contents,the characteristic (pulsed long connection)of WeChat traffic is analyzed,and the classification model is put forward. On the basis of the classification model,the protocol feature of WeChat is analyzed,and the WeChat traffic is identified accurately by matching the effective load. An algorithm to identify the WeChat traffic futher is poposed to classify the WeChat sub⁃business with fine grit. The experiment was per⁃formed to classify the WeChat traffic. The algorithm can identify 52%WeChat sub⁃business with 98%classification accuracy.%以微信流量模型和业务识别为主要研究内容，首先分析了微信流量的特点，即脉冲式长连接的方式，并提出了分类的模型；其次，在分类模型的基础上分析了微信的协议特征，通过匹配有效载荷识别出微信流；同时，提出了进一步识别出微信流量的算法以及细粒度分类微信子业务；最后，进行了实验分类微信流，以98%的分类准确度识别出52%的微信子业务。