Quick traffic classification of BT based on its handshake packets

Abstract

With the popularity of the Internet, the extension speed of network resources always lags behind the demands of network bandwidth of network users. It is not an omnipotent solution to increase the available bandwidth. That is why it is important to analyze, control and manage network traffic accurately nowadays. It is reported that, as one of the prominent occupants, peer-to-peer (P2P) traffic takes up at least 70% of the total bandwidth. And BitTorrent (BT) traffic accounts for the first place among all P2P traffic. In an offline analysis on campus traffic captured in our university, we propose a method based on BT handshake packets which can accurately identify and classify BT traffic at a very low cost before evaluating our proposed method with Deep Packet Inspection (DPI) method whose results show that our method needs a much less classification time and storage space which outperforms the DPI-based BT traffic classification.