A P2P Traffic Detection Method Based on Support Vector Machine

Abstract:

The widely use of P2P (Peer-to-Peer) technology has caused resources take up too much, security risks and other problems, it is necessary to detect and control P2P traffic. After analyzing current P2P detection methods, a new method called TCBDM (Traffic Characters Based Detection Method) is put forward which combines P2P traffic character with support vector machine to detect P2P traffic. By choosing P2P traffic characters which differ from other network traffic, such as Round-Trip Time (RTT), the method creates a SVM classifier, uses a package named LIBSVM to classify P2P traffic in Moore_Set data sets. The result shows that TCBDM can detect P2P traffic effectively; the accuracy could reach 98%.

• Periodical: Applied Mechanics and Materials (Volumes 198-199)
• Pages: 1280-1285
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.198-199.1280
• Online since: September 2012
• Authors: Shang Fu Gong, Juan Chen
• Keywords: P2P, Support Vector Machine (SVM), Traffic Character, Traffic Detection

References

[1] Wang Xiao-wei, A research of P2P traffic identification and detection, [D]. Chongqing: Chongqing University, 2008, p.68–73.

[2] Basher , Mahanti A, Mahanti A, et al. A comparative analysis of web and peer-to-peer traffic[C]. Proceeding of the 17th international conference on World Wide Web. Beijing, China, 2008, 4: 21-25.

DOI: 10.1145/1367497.1367537

[3] Nei Bo. The research on a method of P2P traffic detection based on flow features, [D]. Beijing: Beijing University of Posts and Telecommunication, 2009, p.35–42.

[4] Wu Min, Wang Ru-chuan. P2P Traffic Identification Using Variance Analysis and Support Vector Machine Alagorithm, [J]. Computer Science, 2010, 37(8): 88~113.

[5] Wang Rui. The Study of P2P Traffic Identification Methods, [D]. Changsha : National University of Defense Technology S, 2006, p.21–24.

[6] MOORE AW, ZUEV D. Internet traffic classification using Bayesian analysis techniques[C]. Proceedings of the 2005 ACM SIGMETRICS International Conference on Measurement and Modeling of ComputerSystems. New York: ACM Press, 2005: 50-60.

DOI: 10.1145/1064212.1064220

[7] TAO Xiao-ling, HU Ting. A Network Traffic Classification Method Based on the MIX-Kernel Self-Organizing Maps, [J]. Computer Engineering & Science, 2010, 32(10): 23~25.

DOI: 10.1109/iciss.2010.5657079

[8] HU Ting, WANG Yong, TAO Xiao-ling. Network traffic classification based on hybrid model, [J]. Joumal of Computer Applicatons, 2010, 30(10): 2653~2655.

DOI: 10.3724/sp.j.1087.2010.02653

[9] JIANG Wei, WANG Chun-zhi, LUO Hong-fang, CHEN Hong-wei. Research on a Method of P2P Traffic Detection Based on SVM, [J]. Journal of Hubei University of Technology , 2010, 25(2): 48~50.