基于C4.5决策树的单向P2P流量识别方法研究

发布时间：2018-07-02 23:23

本文选题：C.决策树 + 单向流　；参考：《小型微型计算机系统》2013年02期

【摘要】：P2P流量识别方法已能对基于TCP传输的双向P2P流量进行高准确率的识别,但少有论文研究单向P2P流量的识别方法.针对单向P2P流量的识别难点,提出一种基于C4.5决策树的单向P2P流量识别方法.该方法仅利用10个网络流统计特征,且这些特征可在一个流的前若干个数据包内快速计算完成,并对承载P2P流量的传输层协议具有通用性.实验结果表明,该方法识别准确率高,识别时间短,可用于高速骨干链路P2P流量的实时识别.
[Abstract]:P2P traffic identification method has been able to identify two-way P2P traffic based on TCP with high accuracy, but there are few researches on one-way P2P traffic identification method. Aiming at the difficulty of one-way P2P traffic identification, a one-way P2P traffic identification method based on C4.5 decision tree is proposed. This method only utilizes 10 network flow statistical features, and these features can be quickly calculated in the first several packets of a flow, and are universal to the transport layer protocols carrying P2P traffic. Experimental results show that this method has high recognition accuracy and short recognition time. It can be used to identify P2P traffic in high-speed backbone links in real time.
【作者单位】：国家数字交换系统工程技术研究中心;
【基金】：国家科技支撑计划项目(2011BAH19B01)资助
【分类号】：TP393.06;TP311.13

【相似文献】