Ares Protocol Analysis And Research On Traffic Detection Mechanism

With the birth of computer, only in few decades, the development of computer networkshas made a great change to our society. From the initial terminal network, C/S and B/S modenetworks to today's p2p mode network superimposed with a variety of networks, all those bringus a lot of happiness, at the same time, bring us troubles too. For example, the prosperity of thenetwork communication gives an increasing pressure to the backbone of the computer net, thenetwork operator are under great burden too, especially with the beginning of the p2p modenetwork. In this case, we must monitor the P2P traffic and take appropriate control measures.ARES(Ares galaxy), a popular multifunctional P2P file sharing software in South America,maintains the whole filesharing network by using the DHT network which combined with lotsof super ARES nodes that are running over a few ordinary ARES clients. ARES client caneasily share its pictures, musics and videos to others and even chat with friends online with thehelp of DHT net. All those attract a large number of users, bring a wealth of net resource andoverload the net too. So it is very important to find a wise way to control the ARES trafficwithout affecting users.In this thesis, we analysis its protocol and network topology first, and then identify theARES protocol by using deep packet detecting and dynamic flow detecting, by this way we candirectly get a more accurate flow detecting result by comparing with the results of deep packetdetecting. With these more accurate flow samples, we can also get a much more excellentneural network by automatic learning to replace the old ones, so that we can improve theidentification efficiency and lay the foundation for the flow control.