Research on dynamic network security detection technology based on SDN

Traditional network architecture can no longer meet the development needs of technologies such as cloud computing and big data. SDN network architecture has the characteristics of high openness and programmability, which can quickly respond to changes in business requirements. SDN decouples the control management function of the network from the data forwarding function, simplifies network configuration work based on logic centralized controllers and open programming interfaces, thereby achieving a flat management mode and flexible data forwarding function. The centralized control and programmability of SDN networks also pose significant security threats to SDN. Based on the CSE-CIC-IDS2018 dataset, this article tests the performance of various intelligent learning algorithms, analyzes data flows, distinguishes normal data from abnormal data, improves detection accuracy, and has practical significance for the security defense of SDN networks.