Internet Traffic Classification by Aggregating Correlated Decision Tree Classifier

Peer-to-Peer (P2P) traffic shows a rapid growth in recent times. For efficient malware detection and network traffic management P2P network traffic classification is essential. The existing P2P traffic classification methods includes port-based, signature-based, pattern-based, and statistics based methods. However, none of these methods proved to be effective for the traffic classification in terms of the classification accuracy. This paper proposes a novel classification technique which classifies the internet traffic into P2P and non-P2P traffic with more accuracy and less computational overhead. The proposed classifier is the flow based classifier, that analyses the behavioural patterns utilizing the correlation metric algorithm. The proposed classifier is analyzed for its performance and the results are encouraging.