Internet Traffic Classification Based on Incremental Support Vector Machines
- 162 Downloads
Machine learning methods have been deployed widely in Internet traffic classification, which identify encrypted traffic and proprietary protocols effectively based on statistical features of traffic flows. Among these methods, support vector machines (SVMs) have attracted increasing attention as it achieves the state of art performance in traffic classification compared with other machine learning methods. However, traditional SVMs-based traffic classifier also has its limitations in real application: high training complexity and computation cost on both memory and CPU, which leads to the frequent and timely updating of traffic classifier being impractical. In this paper, incremental SVMs (ISVM) model is first introduced to reduce the high training cost of memory and CPU, and realize traffic classifier’s high-frequency and quick updates Besides, a modified version of ISVM model with attenuation factor, called AISVM, is further proposed to utilize valuable information in the previous training data sets. The experimental results have proved the effectiveness of ISVM and AISVM models in traffic classification.
KeywordsInternet traffic classification Incremental learning Support vector machines Attenuation factor
This work was partly financially supported through grants from the National Natural Science Foundation of China (No. 60903083 and 61502123), Scientific planning issues of education in Heilongjiang Province (No. GBC1211062), and the research fund for the program of new century excellent talents (No. 1155-ncet-008). The authors thank the anonymous reviewers for their helpful suggestions.
- 1.Kim H, Claffy K C, Fomenkov M, et al (2008) Internet traffic classification demystified: myths, caveats, and the best practices[C]//Proceedings of the 2008 ACM CoNEXT conference. ACM, 1-12Google Scholar
- 2.Karagiannis T, Papagiannaki K, Faloutsos M (2005) BLINC: multilevel traffic classification in the dark[C]//ACM SIGCOMM Computer Communication Review. ACM 35(4):229-240Google Scholar
- 3.Nguyen T T T, Armitage G (2008) A survey of techniques for internet traffic classification using machine learning[J]. IEEE Commun Surveys Tutor 10(4):56-76Google Scholar
- 4.Wang Y, Yu S Z (2008) Machine learned real-time traffic classifiers[C]. Intelligent Information Technology Application, 2008. IITA’08. Second International Symposium on IEEE, 3 p 449-454Google Scholar
- 11.Syed NA, Liu H, and Sung KK (1999) Handling concept drifts in incremental learning with support vector machines. In Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM Press, New York, NY, pages 272-276Google Scholar
- 12.A. Moore and K. Papagiannaki. Toward the Accurate Identification of Network Applications. In Proceedings of PAM Workshop, 2005Google Scholar
- 16.Perera P, Tian Y C, Fidge C, et al (2017) A Comparison of Supervised Machine Learning Algorithms for Classification of Communications Network Traffic[C]//International Conference on Neural Information Processing. Springer, Cham, 445-454Google Scholar
- 17.Moore A and Zuev D (2005) Internet traffic classification using Bayesian analysis techniques[C]. ACM SIGMETRICS-2005, Banff, Alberta, p 50-60Google Scholar
- 24.Ruping S. (2001) Incremental learning with support vector machines[C]. Data Mining, 2001. ICDM 2001, Proceedings IEEE International Conference on. IEEE, p 641-642Google Scholar
- 27.Tsai C H, Lin C Y, Lin C J. (2014) Incremental and decremental training for linear classification[C]. Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, p 343-352Google Scholar
- 28.Moore AW and Zuev D (2005) Discriminators for use in flow-based classification [R]. Technical report, Intel Research, CambridgeGoogle Scholar