Network Traffic Monitoring And Analysis Technology Research

In recent years, the application of data is transmitted by network. Network traffic is the main form to bear the network data. Therefore, the identification of network traffic, and extraction and analysis of network data and realization of flow measurement and control become increasingly important. At present, the main goal of many companies is to identify the network traffic more and more, so as to achieve the effective control and the use of network source. While the P2P flow is larger than others in the total flow, and have less network application, they are using in stream media file transfer and timely communication. Therefore, identification of P2P traffic has become the main content analysis of traffic identification.The thesis firstly introduces the traditional traffic identification methods, and introduces the automatic feature recognition based on the traditional method. This method is only suitable for the single flow, and not encrypted data flow analysis, and is not applicable in P2P flow. This paper through the machine learning algorithm carries on the careful analysis to the P2P flow and finding out the characteristics of motion used for the identification of P2P traffic, then using the FCD fuzzy identification P2P network traffic. Experiments show that current data flow recognition detection methods have high accuracy.The innovation point of the thesis is use of FCD fuzzy identification method of P2P traffic identification. And then build log system. Through the analysis of log system, we can apply specifically to generate traffic application. Then we can detect and control of network traffic effectively, and make more rational use of cyber source.