Research On P2P Media Streaming Traffic Identification

In recent years, P2P-based digital media become a public industry. China P2P media streaming industry, represented by PPLive and PPStream, has been having rapid progress, the user scale and the income scale achieved simultaneous growth, while people enjoy the advantage which the distributional pattern brings, it also created tremendous network bandwidth consumption. The research of P2P media streaming traffic identification has important significance in network monitoring, content monitoring and network QoS improvement. While P2P streaming media is a new application, therefore, the current research on this kind traffic identification is still very limited. In view of P2P streaming media software is commercial software, the agreement does not open, in addition to taking the dynamic ports, hidden flows, load encryption technology, the traditional identification methods can not recognized such kind application effectively. These require further more researches on extraction and in-depth analysis of flow characteristics and operational characteristics of P2P streaming media and exploring new ways to improve the recognition accuracy.This thesis takes example by PPStream, which is one of the popular stream media, firstly introduces the recent research situation, and analyzes the advantages and disadvantages of diverse conventional general P2P traffic identification methods, then analyzes P2P streaming media traffic flows characteristics, also researches on traffic identification methods through network traffic monitoring. The thesis especially puts emphasis on distinguishing methods bases on integrated characteristics specifically for diverse P2P media streaming traffic and user behavior, accurately and effectively. Finally, an active achieving P2P streaming media information system is designed and implemented which takes P2P streaming media traffic identification and user behavior identification technology as the core. The system active achieves the P2P steaming media service and user behavior information. Moreover, it could present the characteristic of diverse P2P streaming media traffic and user behavior through the friendly interface, directly and clearly.