A method of identifying P2P streaming media traffic flow in the unknown, is the first of a large number of statistics contains many applications of P2P streaming media data stream, extracting the packet payload length as a parameter, derived from multi class samples in accordance with the application of discrete distribution characteristics of interval; UDP flow of the unknown flow quantity in the transport layer based on the data flow as input unit, wherein each packet and obtaining its extraction load length, according to the application of sample prior probability of the packet belongs to each type of application value, which belongs to the application of each type of calculation formula based on Bayesian posterior probability values, and taking the maximum value; the definition of said data stream reliability belongs to a kind of applications the credibility of the value, the maximum and the average value was selected by iteration combined with posterior probability, as the data stream current reliability value; the current value and the reliability of data flow By comparing the default decision threshold, the application category that determines the ownership of the data stream is determined. The invention has the advantages of fast processing speed, small storage space and wide application range.
【技术实现步骤摘要】
【技术保护点】
一种在未知流量中识别点对点Ｐ２Ｐ流媒体流量的方法，其特征在于：统计大量的包含多类应用的Ｐ２Ｐ流媒体数据流，提取其中分组的负载长度作为参量，推导得出其符合区间离散分布特性的多类应用样本；对于未知流量，将传输层的用户报文协议ＵＤＰ流量以数据流为单位输入，提取该数据流中的各个分组而分别获知其负载长度，根据应用样本得出该分组属于每类应用的先验概率值，再结合贝叶斯公式计算其属于每类应用的后验概率值，并从中得出其后验概率的最大值；接着，定义数据流的信度值，用于表示该数据流属于某一类应用的可信度，再结合后验概率的最大值和通过迭代选取的平均值，作为该数据流的当前信度值；再将该数据流的当前信度值和预设的判定阈值相比较，就判定得出该数据流所归属的应用类别。
【技术特征摘要】
还没有人留言评论。发表了对其他浏览者有用的留言会获得科技券。