Research on Detection Method of Abnormal Traffic in SDN

摘要

Compared with traditional network, the network architecture and equipment function of SDN have changed dramatically. Thus it is necessary to research more targeted network security strategies. Abnormal traffic detection is the foundation of intrusion detection and intrusion prevention. For this reason, This paper proposes a specific abnormal flow detection method aimed at SDN. The method makes full use of flow-table in SDN switch to extract the features of abnormal flows, and applies information entropy to process non-numerical features of a flow into numerical features. Finally, a BP neural network model previously trained by these numerical features are used for abnormal flows detection. The contrast experiment results show that, this method can detect abnormal traffic in SDN effectively.