P2P traffic identification research based on the SVM

This paper first introduces the advantages and disadvantages of all kinds of P2P traffic identification method especially machine learning traffic identification method, and then puts forward a method of P2P traffic identification model based on SVM. We select three characteristics including the change of the mean square value of P2P traffic data packet size, the time of average flow duration and the ratio of the IP address number and port number to identify the network flow. The experimental results show that the method can effectively detect the P2P traffic of network flow.