A P2P Traffic Identification Approach Based on SVM and BFA

Abstract

Nowadays new peer to peer (P2P) traffic with dynamic port and encrypted technology makes the identification of P2P traffic become more and more difficult. As one of the optimal classifiers, support vector machine (SVM) has special advantages with avoiding local optimum, overcoming dimension disaster, resolving small samples and high dimension for P2P classification problems. However, to employ SVM, the parameters selection of SVM should be considered and thus some optimization methods have been put forward to deal with it, still, it is not fully solved. Hence, in the paper, a peer to peer traffic identification approach based on support vector machine and bacterial foraging algorithm is proposed for better identification of P2P network traffic. First, the best parameters for SVM are tuned with bacterial foraging algorithm. Subsequently, SVM set with the best parameters is used to identify P2P traffic. Finally, experimental results show the proposed approach can effectively improve the accuracy of P2P network traffic identification.

DOI : http://dx.doi.org/10.11591/telkomnika.v12i4.4736