Research On Micro - Credit Flow Model And Service Identification Method

With the continuous development of Internet technology, network communication quality becomes more and more important, as a carrier of Internet communications, network traffic attracts closed attention. Traffic classification is a preliminary way of the traffic analysis and an indispensable process of understanding internet traffic. As the new application and requirement springs up, the growth of network traffic becomes diversified, and gives internet management and operation enormous pressure and challenges.Real-time network traffic classification is conducive to help internet service providers to understand the network status, and plays a decisive role in the optimization of network operation and management. Internet service providers can statistic traffic and forecasts the trend of network business, optimize network infrastructure by the classification of traffic. According to classification, analysis the different application in different kinds, can well realize the deployment of quality of service(QoS). Different services for different applications can avoid network congestion, and ensuring the quality of critical services for efficient and smooth streaming. In addition, in the aspect of network security, traffic classification is the core of the intrusion detection system, and take the necessary measures in time by finding the unknown or anomaly traffic through traffic classification.In recent years, WeChat has a rapid development and promotes mobile operators changing the business model, has received the widespread attention. Based on WeChat traffic model and identification as the main research content, the paper firstly analyzed the flow characteristic of WeChat, concluded that the traffic is a pulsed long-live connection, and put forward the classification model; Second, on the basis of the classification model, has analysized the protocol signature, and accurately identify the WeChat traffic using deep packet inspection; Then, futher poposed an identification algorithm of WeChat flow to mark out different child business of WeChat; Finally, experiment has done with WeChat flow with 98% of the classification accuracy of identified 52% the business.