高速网络下P2P流量识别研究

doi:10.3969/j.issn.1671-1122.2015.05.011

摘要/Abstract

摘要：

网络流量分类是指将混有各种应用的流量按照流量所使用的协议进行分类。网络流量分类一直是各界共同关注的热点之一。研究网络流量分类可以为设计下一代高性能网络协议提供基础,可以为网络管理和网络流量调度提供依据,可以为网络攻击防护和流量清洗提供支撑。文章分析了当今主流的网络流量识别方法,归纳总结了流量分类技术的发展现状和研究成果。文章针对P2P流量迅猛增长的现状和高速网络下数据流量的特性,重点研究高速网络下P2P流量的二元分类方法,首先采用基于传输层行为的流量识别方法将部分样本数据流分为P2P流量和非P2P流量,并根据分类结果动态生成有标记的特征训练集,有效避免了网络数据流发生变化后导致分类算法所使用的样本特征集不准确的情况;然后提出一种基于C4.5决策树算法的P2P流量识别方法,该方法只需计算一个数据流的前若干个数据包即可完成流量识别,无需考虑数据流的单双向和数据加密等问题。实验表明该方法识别准确率高,计算复杂度低,适用于高速网络下的流量识别。

关键词: 高速网络, P2P, 流量识别

Abstract:

Network traffic classification refers to classify the flow which mixed with a variety of applications in accordance with the protocol which flow used. Network traffic classification has been one of the hot spot in all walks of life. Research on network traffic classification can provide the basis to design the next generation of high performance network protocol, can provide the gist for network management and network traffic scheduling, can provide support for network attack protection and traffic cleaning. This paper analyzes the network traffic identification methods in nowadays mainstream sand summarized the flow classification technology development present situation and research results. According to current situation of rapid growth of P2P traffic and the characteristics of high-speed network traffic, focuses on the binary classification method of P2P traffic which under the high-speed networks. In this paper, data stream can be divided into P2P traffic and the normal network traffic by the traffic identification methods based on the behavior of the transport layer firstly, and dynamically generate marked characteristic of the training set according to the results of the classification. Effectively avoids the sample set which is used by the classification algorithm is not accurate because of the network data flow changed. And then put forward a kind of P2P traffic identification method based on C4.5 decision tree, this method only need to calculate several packets which in a data flow, and then the network traffic identification is completed. Don’t need to concern of single or double direction of the data flow, data encryption, etc. Experiments show that the recognition accuracy of this method is high; the computational complexity is low, suitable for high-speed network traffic identification.

Key words: high speed network, P2P, traffic identification

中图分类号:

TP309

穆筝, 吴进, 许书娟. 高速网络下P2P流量识别研究[J]. 信息网络安全, 2015, 15(5): 69-76.

Zheng MU, Jin WU, Shu-juan XU. Research on P2P Traffic Identification Under the High Speed Network[J]. Netinfo Security, 2015, 15(5): 69-76.

图/表 4

参考文献 37

[1]	WANG J, ZHOU Y, YANG Y, et al.Classify the majority of the total bytes on the Internet[C]//Information Processing (ISIP), 2008 International Symposiums on, Moscow, 2008.
[2]	KIM M S, KANG H J, HONG J W.Towards peer-to-peer traffic analysis using flows[C]//Self-Managing Distributed Systems, LNCS 2867. Berlin: Springer, 2004: 55-67.
[3]	鲁刚,张宏莉,叶麟. P2P流量识别[J]. 软件学报,2011,22(6):1281-1298.
[4]	KARAGIANNIS T, BROIDO A, FALOUTSOS M.Transport layer identification of P2P traffic[C]//New York, NY, USA: ACM, 2004: 121-134.
[5]	HAFFNER P, SEN S, SPATSCHECK O, et al.ACAS: Automated construction of application signatures[C]//SIGCOMM’05 Workshops, New York, 2005: 197-202.
[6]	BASET S, SCHUNZRINNE H G.An analysis of the Skype peer-to-peer Internet telephony protocol[C]//Proceedings of the 25th IEEE International Conference on Computer Communications, USA: INFOCOM, 2006: 1-11.
[7]	袁雪美,王晖. P2P流量识别技术综述[J]. 计算机应用,2009,29(5):12-13.
[8]	MARIOS I, PAPPU P, FALOUTSOS M. Network Traffic Analysis using Traffic Dispersion Graphs (TDGs):Techniques and Hardware Implementation[J]. Technical Report, 2007, UCR-CS-2007-05001: 15-18.
[9]	GU CHENGJIE, ZHANG SHUNYI, XUE XIAOZHEN.Encrypted Internet Traffic Classification Method based on Host Behavior[J]. Internet Journal of Digital Content Technology and its Applications, 2011, 5(3): 167-174.
[10]	BRANCH P A, HEYDE A, ARMITAGE G J.Rapid identification of Skype traffic flows[C]//NOSSDAV’09. USA: Virginia, 2009: 91-96.
[11]	FREIRE E P, ARTUR Z, RONALDO M.On Metrics to Distinguish Skype flows from HTTP traffic[J]. Journal of Network and Systems Management, 2009, 17(1-2): 57-66.
[12]	EHLERT S, SANDRINE PETGAND.Analysis and Signature of Skype VoIP Session Traffic[C]//Fraunhofer FOKUS Technical Report NGNI-SKYPE-06b, 2006:14-15.
[13]	IDREESS F, UZMA A K.A generic technique for Voice over Internet Protocol(VoIP) traffic detection[J]. IJCSNS International Journal of Computer Science and Network Security, 2008, 8(2): 52-59.
[14]	ROUGHAN M,SEN S,SPATSCHECK O.Class-of-service mapping for QoS: a statistical signature-based approach to IP traffic classification[C]//Proceedings of the 4th ACM SIG- COMM conference on Internet measurement. Taormina, Sicily, Italy: ACM, 2004: 135-148.
[15]	CONSTANTINOU F, MAVROMMATIS P.Identifying Known and Unknown Peer-to-Peer Traffic[J]. NCA, 2006: 93-102.
[16]	SEN S, WANG J.Analyzing peer-to-peer traffic across large networks[J]. IEEE-ACM TRANSACTIONS ON NETWORKING, 2004, 12(2): 219-232.
[17]	KARAGIANNIS T,PAPAGIANNAKI K,FALOUTSOS M.BLINC: multilevel traffic classification in the dark[C]// SIGCOMM’05, New York, 2005: 229-240.
[18]	ALSHAMMARI R, LICHODZIJEWSKI P, HEYWOOD M, et al.Classifying SSH Encrypted Traffic with Minimum Packet Header Features using Genetic Programming[C]//Defense applications of computational intelligence workshop, USA: GECCO, 2009: 2539-2546.
[19]	THANH V V, FUKUDA K, CHAN H N, et al.A study on P2P traffic characteristics evaluation in the WIDE backbone[C]//The 6th ICITA, USA, 2009: 286-291.
[20]	ESTEVES A F, INACIO P, PEREIRA M, et al.On-line Detection of Encrypted Traffic Generated by Mesh-Based Peer-to-Peer Live Streaming Applications: The Case of GoalBit[C]//IEEE 10th International Symposium on Network Computing and Applications, 2011: 223-228.
[21]	ROSSI D, VALENTI S.Fine-grained traffic classification with Netflow data[C]//IWCMC, France, 2010: 479-483.
[22]	Gomes J V P, INACIO P, FREIER M, et al. Analysis of Peer-to-Peer Traffic Using a Behavioural Method Based on Entropy[C]//The 27th IEEE IPCCC, USA, 2008: 201-208.
[23]	鲁刚,张宏莉. P2P流量分类研究[J]. 智能计算机与应用,2013,3(3):2-3.
[24]	MUNZ G, HECKMULLER S, BRAUN L, et al.Improving Markov-based TCP Traffic Classification[C]//17th GI/ITC Conference on Communication in Distributed Systems, Viet nam, 2010: 61-72.
[25]	MOORE A W, ZUEV D.Internet Traffic Classification Using Bayesian Analysis Techniques[C]//SIGMETRICS’05, Canada: ACM, 2005: 50-60.
[26]	ERMAN J, ARLITT M F, MAHANTI A.Traffic classification using clustering algorithms[J]. MineNet, 2006: 281-286.
[27]	ERMAN J, MAHANTI A, ARLITT M F.Byte me: a case for byte accuracy in traffic classification[J]. MineNet, 2007: 35-38.
[28]	ABI P S.Efficient Detection of VoIP calls Hidden in Web Traffic[J]. IJTES, 2011, 2(1): 1-4.
[29]	AULD T, MOORE A W, GULL S F.Bayesian Neural networks for Internet Traffic Classification[J].IEEE Transactions on neural networks, 2007, 18(1): 223-239.
[30]	YUAN R X, LI Z, GUAN X H, et al.An SVM-based machine learning method for accurate internet traffic classification[J]. Information Systems Frontiers, 2010, 12(2): 149-156.
[31]	戴磊. 流分类技术研究[D]. 北京:中科院计算所,2009.
[32]	SALGARELLI L, GRINGOLI F, KARAGIANNIS T.Comparing Traffic Classifiers[J]. ACM SIGCOMM Computer Communication Review, 2007, 37(3): 65-68.
[33]	KIM H, CLAFFY K, FOMENKOV M.Internet traffic Classification Demystified: Myths, Caveats and the Best Practices[C]//CoNEXT 2008, Madrid: ACM 2008: 10-12.
[34]	WILLIAMS N, ZANDER S, ARMITAGE G J.A preliminary performance comparison of five machine learning algorithms for practical IP traffic flow classification[J]. Computer Communication Review, 2006, 36(5): 5-16.
[35]	KIM H, BARMAN D, KWON T T, et al.NeTraMark: a network traffic classification benchmark[J]. Computer Communication Review, 2011, 41(1): 22-30.
[36]	COLLINS M, REITER M.Finding peer-to-peer file-sharing using coarse network behaviors[C]//The European Symposium on Research in Computer Security, Germany, 2006: 73-75.
[37]	DAINOTTI A, PESCAPE A, ROSSI P S.Internet traffic modeling by means of Hidden Markov Models[J]. Computer Networks, 2008, (52): 2645-2662.