Real-Time Identification Research of Unstructured P2P Multicast Video Streaming
Chaobin Liu, Jie He, Qiang Guo · 2012
Many potential safety problems result from the rapid development of P2P IPTV business, and the foundation for effectively managing the business is to accurately identify unstructured P2P multicast video streaming. In this paper, an identification method is proposed which is based on support vector machines. The network traffic is successively separated by flow features and behavior features, and finally the applications of unstructured P2P multicast video streaming could be identified. The method not only could adapt to the continuous changes of network, but also could identify the known and unknown flow of unstructured P2P multicast video streaming online. The experiments show that the average identification accuracy of the method is 90.9%, and the identification time is about 5 minutes.