Skype Traffic Identification Based on Trends-Aware Protocol Fingerprints

Abstract

The P2P technology consumes the largest proportion of network traffic and is usually encrypted, which is lack of supervision. Accurate and rapid identification of encrypted P2P traffic, represented by the famous Skype, is of great significance to improve the network quality of service and enhance security control. In this paper, a trends-aware protocol fingerprints model is proposed based on the statistical signatures of signaling interactions and content transfer phase of Skype. The proposed method can sense traffic trends by trends-aware weighting functions and identify Skype traffic with anomaly scores in real-time. Experimental results show that the precision and real-time performances of the proposed algorithm is better than several state-of-art encrypted traffic identification methods, such as the protocol fingerprints and C4.5 algorithm.