The identification for P2P Thunder traffic based on deep flow identification

Abstract

Network traffic analysis and identification play important roles in network traffic monitoring. The network applications are the root causes to generate network traffic and network behavior. Network traffic analysis and identification is the basis of other network problems, which provides ISP an effective basis to control and distinguish the network traffic. With the popularity of Internet, "download" has become one of the most important parts of the domestic Internet users. As more and more resources, there is more and more enthusiastic discussion of the download tools. This paper takes one of the most popular download applications Thunder (also called Xunlei) for example, indicating the significance of the flow-based statistical network traffic classification and identification. Through the characteristics extraction, we can identify the Thunder traffic with high accuracy, and at the same time, the time cost in the experiment is reduced a lot.