Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-08-13 Cooperative journals: 《计算机应用研究》
Abstract: In order to effectively supervise the network, quickly and accurately identify the peer-to-peer flow, by analyzing the interaction and behavior characteristics between nodes and nodes, nodes and links in Peer-to-peer network traffic, a method of Peer-to-peer traffic recognition based on network behavior features is proposed by combining clustering method with flow propagation graph method. Firstly, the flow rate of different kinds of network is collected by collecting packet level and flow level statistic feature of network flow, and then the Peer-to-peer flow is identified by using traffic graph. The experimental results show that the proposed method can effectively identify Peer-to-peer network application traffic in backbone network data, and the f1-measure reaches over 95%.