Skype traffic identification based SVM using optimized feature set

Skype traffic recognition is a challenging problem due to the encryption and dynamic port number. Accuracy and timely traffic classification is critical in network security monitoring and traffic engineering. In this paper, we propose an online recognition method based on SVM (support vector machine) machine learning method. As the feature set is optimized instead of redundant, our method is able to compute faster and more accuracy. Experimental results on Collage campus data sets show that our method performs better on both speed and efficiency. Moreover, the robustness of our method is demonstrated on the other non-Skype traffic such as MSN (Microsoft Service Network), PPLive (Peer to Peer LIVE) application.