Traffic classification method of P2P streaming application based on sampling technology

With the continuous development of high-speed network technology, this paper presents a traffic classification method of P2P streaming application based on sampling technology. By comparing the statistics data before and after the sampling, we find a group of statistical flow characteristics which are easy to be acquired and implemented. Using these characteristics and classification method proposed by this paper, we obtained a higher recognition rate of P2P streaming application. More importantly, traffic of P2P streaming with different sampling rate could be effectively identified by using unified classification model. The accuracy rate of classification is more than 92% when sampling rate was not less than 0.01. It is qualified for highspeed network measurement.