The Traffic Identification Technique Research On Audio And Video Across Large Networks

With the great development of the internet, a great deal of new applications are emerging, such as P2P applications, IPTV, VoIP and Online Games. The proportion of traffic stream has a greatly change.The increasing traffic of audio and video brings some impact on the network, which greatly increases the burden on the network and causes network congestion. Detecting and analyzing the traffic of audio and video becomes an important part for the ISP to manage the network. Meanwhile, in order to improve the quality of service of multimedia we also need to detect and analyze the traffic. The accurate and rapid identification is the base of traffic detection and management, therefore, the traffic of audio and video becomes the focus of our research.There are about four majority methods for traffic identification. The port based traffic classification, the characteristic of traffic based classification, the protocol based classification and the machine learning based classification. But each method has their own advantages and disadvantages, and has limits to process the traffic of audio and video. In order to process the traffic of audio and video much better, we did some research on the classification of audio and video, analyzed the traffic identification methods. By analyzing the advantages and disadvantages of these methods we proposed a multi-level traffic identification based on flow, it used the TCP and UDP flow to identify the traffic of audio and video, we put forward a new UDP flow detection strategy, reduced the consumption of resource and avoided processing every packet in the great traffic. The multi-level identification technique combined the advantages of some traffic identification methods to improve the accuracy of the result. Finally, we proved the effective and accuracy of our method by experimental traffic, and analyzed the real traffic.Flow-based multi-level traffic identification method can provide fine-grained result, and provides a general method to the traffic of audio and video. It is more convenient for detecting and managing the traffic and provides reference for analyzing the quality of service of audio and video.