中文摘要：

使用流量特征的方法可有效实现对未知P2P流媒体流量及加密后流量的辨识，具有可扩展性．总结出多源型P2P流媒体系统区别于P2P文件下载系统的资源高动态性、超细粒度的分块和由此引起的高频度信息交换及调度的独特特性，因此P2P流媒体节点的流量中应具有交换信息的通讯包发送频率高的特征，从而提出通过计算节点入流量的平均连接信息数据比来识别多发送节点方式下的P2P流媒体流量的观点，给出信息数据比的定义及测量模型，并对测度值进行了推导和实验，证明了这些测度值的可用性．

英文摘要：

The method for identifying P2P media streaming by traffic features is scalable and suitable for any unknown and encrypted traffic. Some characteristics of P2P media streaming distinguished from P2P file-sharing traffic are provided in this paper, and based on it, the conclusion is gained that the information packets are sent in much higher frequency in P2P media streaming traffic than the file-sharing traffic. Then,the method is proposed that the average value of the ratio of the time in which information packets are transferred to the one in which all data packets transferred of all flows from uploading peers-to downloading peer can be used as a metric for identification. The thresholds of this metric are discussed under the measuring model. The experiment shows the feasibility of this metric as well as its thresholds.