摘要: 针对VoIP加密负载流量识别的难题,提出一种基于UDP统计指印混合模型的VoIP流量识别方法,以提高VoIP流量的识别精度和分类稳定性。该模型改进了统计指印模型中基于单一的网络流相异度来判定流量类别的方法,将UDP流的统计特征与网络流的统计指印相异度结合以共同训练一个支持向量机分类模型,把基于分类阈值点的分类转换到基于多维特征的高维空间中的分类面的分类,综合运用包层次和流层次统计特征,降低了因网络不稳定造成的统计特征偏差对分类模型精确度的影响。实验结果表明,该模型对VoIP流量的分类精确度达到97%以上,与统计指印模型和支持向量机模型相比分类稳定性更好。
[1] Salman A.Baset,Henning Schulzrinne.An Analysis of theSkype Peer-to-Peer Internet Telephony Protocol[C]∥Procee-dings of the 2006IEEE Infocom, IEEE’06.Barcelona,Spain,Apr.2006 [2] Moore A W,Zuev D.Internet Traffic Classification using Bayesian Analysis Techniques[C]∥Proceedings of the 2005ACM SIGMETRICS Conference on Measurement and Modeling of Computer Systems.New York,USA:ACM,2005:50-60 [3] 徐鹏,刘琼,林森.基于支持向量机的Internet流量分类研究[J].计算机研究与发展,2009,46(3):407-414 [4] Crotti M,Dusi M.Traffic Classification through Simple Statistical Fingerprinting[J].ACM SIGCOMM Computer Communication Review,2007,7(1):5-16 [5] Bonfiglio D,Mellia M,Meo M.Revealing Skype Traffic:When Randomness Plays with You[C]∥Proceedings of 2007ACM SIGCOMM Computer Communication Review.New York,USA:ACM,2007:37-48 [6] 张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-42 [7] LI Wei,Canini M,Moore A W.Efficient Application Identification and the Temporal and Spatial Stability of Classification Schema[J].Computer Networks,2009,3(6):790-809 [8] Company of MathWorks.MATLAB[EB/OL].http://www.mathworks.cn/products/matlab/,2011-06-02 [9] Joachims T.SVM-Light[EB/OL].http://svmlight.joachims.org/,2011-06-02 [10] Yu Lei,Liu Huan.Feature selection for high-dimensional data:A fast correlationbasedfilter solution[C]∥Proceedings of the 20th International Conference on Machine Learning(ICML’03).2003 [11] COMPANY of SOURCEFORGE.L7-filter[EB/OL].http://l7-filter.sourceforge.net/.2011-06-02 [12] 刘琼,刘珍,黄敏.基于机器学习的IP流量分类研究[J].计算机科学,2010,7(12):35-40 |
No related articles found! |
|