Identify And Research Based On The Behavioral Characteristics Of P2P

As Internet technology advances, the P2P of this technology has been rapid dev-eloppment, according to the data show that the Internet,60%-80%of the data stream to a P2P service, in some areas such as schools:P2P data stream bandwidth has seriously interfered with the normal teaching order, how to effectively manage P2P data stream has become urgent needs and research focus. But the management is to correctly ident-ify the data flow on the basis of development, the article is in this context. P2P recognition technology from the early port number to identify, to the later app-lication layer signature matching to identify, to the use of P2P traffic characteristics to identify, these technologies are in constant development.The port uses a random port in the new generation of P2P applications, it is not simple port number to effectively distinguish between data flow and data transmission is encrypted, not at liberty to make the signature match. Identify P2P data streams is becoming increasingly difficult; how to find an effective and accurate identification of an urgent request. This paper analyzes the P2P traffic characteristics and node characteristics, combined with the BP neural network recognition algorithm, to identify P2P data stream, and on this basis, do the following work:1..Analysis of several typical P2P identification method works, and compare the advantages and disadvantages in the recognition process.2. Analysis of the current domestic popular several P2P software (Thunder, eMule, QQ), the communication process through the captured packets to understand the characteristics of P2P software, to lay the theoretical basis of analysis of P2P traffic characteristics.3. By wireshark P2P and non P2P software to capture one minute of pure data flow analysis of flow characteristics and node characteristics, and characteristics as the input parameters of the neural network with BP algorithm simulation to verify the validity of the parameters in the recognition process. 4. Parameters of these characteristics apply to the snort intrusion detection system, write the rules of the identification of P2P; and rules of application on the host to capture the data stream.