PEER-TO-PEER TRAFFIC IDENTIFICATION

Abstract

The term peer-to-peer refers to sharing of computer resources through direct exchange between computers and peer-to-peer network is a network of personal computers which permits sharing of specific files and folders with everyone or with selected users. Sharing content and finding content are two key functions of peer-to peer systems. Sharing remains the dominant P2P application on the internet, allowing users to easily contribute, search and obtain content so, peer-to-peer traffic is an important class of service. Accurate identification of this traffic is a challenging problem. Port and Payload based techniques are traditional methods of traffic identification. Many peer-to-peer applications use random ports so port based detection does not remain effective. Payload based analysis, though, gives accurate results but has limitations. It is required to pre-determine the signatures and match signatures to identify the traffic. In order to overcome the drawbacks of the traditional identification methods, non-payload based method has been introduced that examines transport layer information and statistics to identify peer-to-peer traffic patterns. It does not deal with user payload and does not require predetermination of any kind as in Payload detection method. The proposed method uses heuristics to implement this technique. It rejects the traffic that show behaviors similar P2P according to applied heuristics but are actually non peer-to-peer.