Research Of Identifying P2P Traffic Based On P2P Traffic Characters

Investigations show that P2P has silently become the dominant force in internet applications, making up 60 to 80 percent of its total volume. While P2P Traffic has strongly promoted the development of Internet, it has also brought about such problems as network congestion owing to resources are over occupied and potential security concern, thus it hampers normal practice of network business. Therefore, relevant P2P traffic control and management methods are required to ensure normal and orderly function of the network. But with the advent of data encryption and dynamic port, P2P traffic identification now faces enormous challenges.This thesis briefly introduced current main P2P traffic identification techniques and focuses on techniques based upon P2P traffic characteristics.Then makes a heavy weight on the analysis about the reasons of low efficiency of the technique based upon P2P traffic characteristics and comes up with their corresponding solutions. Afterwards, regression analysis are taken between several P2P traffic characteristics and their possibility to be identified as P2P . Through detailed analysis,the thesis find out that the number of hosts which a P2P host link with,has a good degree of differentitation only if the number is less than 5 or more than 20 . Furthermore,the characteristic about listening port of P2P hosts has a better degree of differentitation if the number of links which is less than 5 , has been exclueded. Finally, the thesis introduce the latest research about P2P traffic identification and novel ways to identify P2P traffic beyond the techniques discussed above.