A study on SET-based multi-stage identification method for peer-to-peer application

Abstract

As the Internet is rapidly evolving from best-effort network to Quality of Service (QoS) network, the important task is to provide different QoS guarantees to the traffic from different services in the Broadband Convergence Network (BcN). An essential functionality for BcN would be the ability to identify applications for various purposes such as network management, QoS provisioning and access control. One of the main problems with the application identification is the traffic dynamics of the Internet``s dominant applications. There is a strong need for both fast and accurate identification method in QoS network First of all, this thesis focuses on the properties of Peer-to-Peer network and applications. Next, it explains why current P2P traffic analysis is very difficult to identify these applications. To solve the problems, it insists that SET-based identification method in flow-level is a practical way for better accuracy. After that, this thesis proposes a multi-stage P2P identification method. This approach has three different analyzers: Port analyzer, Signature analyzer and SET analyzer. Port-based matching and Signature-based matching ware developed for fast and reliable identification. SET-based matching was developed to identify modified applications which users changed their default service port numbers and hide their protocols. Finally, we will show the result of P2P traffic analysis using the real P2P traffic.