关键词: 对等流量, 流量特征, 流身份识别, 网络地址翻译器, 网络拓扑

Abstract: Mixed flows retransmitted by Network Address Translator (NAT) are similar to Peer-to-Peer (P2P) flows in terms of traffic characteristics. The measurement result shows that without flow identification the TLI method will have high false positives and false negatives. In order to solve this problem, a new P2P traffic identification method based on flow classification was proposed. Firstly, flow classification was done by analyzing IP identification time series. Based on classification results, P2P traffic was identified by checking its traffic characteristics. The proposed method was put to test through popular P2P application. The result shows that it can reduce the false positives and false negatives widely, so new P2P traffic classification method has better application value.

Key words: Peer-to-Peer (P2P) traffic, traffic characteristic, flow identification, Network Address Translator (NAT), network topology