Auto Adaptive Identification Algorithm Based on Network Traffic Flow

Traffic identification is a key task for any Internet Service Provider (ISP) or network administrator. Machine learning method is an important researchmethod on traffic identification, while impact of the asymmetry router on the  traffic identification is considered, so this paper analyzes the impact of asymmetry routing on traffic identification, and proposes an effective method to decrease the impact, and experimental results show the auto adaptive algorithm can improve the traffic identification.