论文部分内容阅读
对等体网络P2P(Peer-to-Peer)应用系统中对等体主机的行为特征与P2P业务流量特征多样化、复杂化,使得单纯利用一种典型特征的P2P流量分类技术的识别精度不高。文中提出了一种新的P2P流量多阶段识别方法,该方法根据P2P应用流量的一系列固有特征,可以从聚合网络流中识别P2P流量。通过实验表明,文中所提出的方法P2P流识别精度可达99.7%,同时错误分类精度0.3%。
Peer-to-Peer (P2P) Peer-to-Peer Application System Peer-to-peer application behavior and P2P traffic characteristics are diversified and complicated, making the classification accuracy of P2P traffic classification technology, which uses only one typical feature, not high . In this paper, a new multi-stage P2P traffic identification method is proposed. According to a series of inherent features of P2P application traffic, this method can identify P2P traffic from aggregated network flows. Experiments show that the method proposed in this paper can achieve 99.7% recognition accuracy and 0.3% error classification accuracy.