Intrusion detection model for software-defined networking and experimental validations

With the continuous development of software-defined networking (SDN) technology, SDN also faces many network security threats. Therefore, timely detection of various attack problems faced by SDN is crucial to solve the SDN network security problems. To address this issue, try to study the intrusion detection model of SDN to screen the intrusion data traffic of SDN. Firstly, on the basis of improving the searching ability of seagull optimization algorithm, the feature selection method based on the improved seagull optimization algorithm is designed to reduce the dimension of the dataset, and then the SVM parameter optimization method based on the improved seagull optimization algorithm is designed to optimize the SVM model. Finally, the proposed two methods are used to construct the SDN intrusion detection model. The experimental accuracy of feature selection on NSL-KDD dataset is 0.945, the experimental accuracy of parameter optimization on UCI dataset is above 0.82, and the accuracy of SDN intrusion detection on InSDN dataset is 0.954. The experimental validation results show that the proposed SDN intrusion detection model can effectively detect the intrusion traffic of SDN network and provide better security guarantee for SDN network.