Online NetFPGA decision tree statistical traffic classifier

摘要: Classifying online network traffic is becoming critical in management and security. Recently, new classification methods based on analysis of statistical features transport layer have been proposed. While these address the limitations port payload classification, current software-based solutions are not fast enough to deal with today's high-speed networks. In this paper, we propose an classifier using C4.5 machine learning algorithm running NetFPGA platform. Our constructed by adding three main modules reference switch design; a Netflow module, feature extractor search tree classifier. The proposed able classify input traffics at maximum line speed platform, i.e. 8Gbps without any packet loss. method first few packets flow. flow classified just micro seconds after receiving desired number packets.