Environment-Independent Online Real-Time Traffic Identification

Abstract

In the current Internet, there is a real need for the online real-time traffic identification technique to provide different services for real-time and bulk applications. Previously, it is easy to identify real-time traffic by checking the protocol/port number in IP header, however, it becomes more difficult due to the existence of real-time traffic over TCP connection, P2P and VPN. Previously, we have proposed the online identification method based on flow statistics without checking the protocol/port number to solve these problems. However, this technique performance is unstable due to environment dependency. In this paper, at first, we reanalyze the characteristics of bulk and streaming traffic flows, which shows that the packet arrival interval varies significantly among high-bitrate, low- bitrate and bulk flows. Second, we propose a new identification method without using a fixed threshold depending on network environment. Finally, testing shows that its identification accuracy is higher than that of a previous method, which recognizes only two types of flows. It also shows that the improved method is robust against differences in the network environment.