| 基于PCA的对等网络流量时空特性监测 |
| |
| 引用本文： | 张文铸,刘佳,袁坚,张林,山秀明.基于PCA的对等网络流量时空特性监测[J].清华大学学报(自然科学版),2010(4). |
| |
| 作者姓名： | 张文铸 刘佳 袁坚 张林 山秀明 |
| |
| 作者单位： | 清华大学电子工程系; |
| |
| 基金项目： | 国家自然科学基金资助项目(60674048,60672107); 国家“九七三”重点基础研究项目(2007CB307100) |
| |
| 摘 要： | 在对等网络中,节点可以动态地进入和离开网络,增加了业务监测的难度。该文提出了采用主成分分析(PCA)方法检测并分析对等网络流量特征,解释了流量协方差矩阵最大特征值和最大特征矢量的物理意义,同时定义权重矢量作为流量观测指标,利用流量的协方差矩阵最大特征值和特征向量给出了全网的P2P流量时间和空间的动态特性。仿真结果表明,该方法能够很好地识别P2P流量的动态时间和空间特性,为互联网流量检测提供了一种有效方法。
|
| 关 键 词： | 对等网络 流量监测 主成分分析 协方差矩阵 网络测量 |
PCA based approach for monitoring the spatial-and-temporal characteristics of P2P traffic |
| |
| ZHANGWenzhu,LIUJia,YUANJian,ZHANGLin,SHANXiuming.PCA based approach for monitoring the spatial-and-temporal characteristics of P2P traffic[J].Journal of Tsinghua University(Science and Technology),2010(4). |
| |
| Authors: | ZHANGWenzhu LIUJia YUANJian ZHANGLin SHANXiuming |
| |
| Institution: | ZHANG+Wenzhu,LIU+Jia,YUAN+Jian,ZHANG+Lin,SHAN+Xiuming(Department+of+Electronic+Engineering,Tsinghua+University,Beijing+100084,China) |
| |
| Abstract: | Monitoring P2P traffic is difficult due to the peers in P2P networks joining and leaving dynamically. This paper presents a method based on principal components analysis (PCA) to monitor the spatial-and-temporal characteristics of P2P traffic. The meaning of the largest eigenvalue and the corresponding eigenvector is explained by the PCA theory,which are calculated from the cross-correlation matrix of the flow data. A weight vector is also defined to monitor traffic patterns. Simulation results show that th... |
| |
| Keywords: | peer-to-peer traffic monitoring principal components analysis cross-correlation matrix network measurement |
| 本文献已被 CNKI 等数据库收录！ |
|
