DDoS Attack Detection and Classification Based on Hybrid Deep Learning Model in SDN

The Industrial Internet of Things (IIoT) involves numerous industrial systems and equipment. Due to challenges in node data processing, a software-defined network (SDN) can improve IIoT control and management. However, SDN’s centralized control can be susceptible to distributed denial of service (DDoS) attacks. Existing deep learning detection models are often complex and computationally demanding when handling high-dimensional datasets. This paper presents a solution for DDoS detection in SDN-enabled IIOT by selecting optimal features with a random forest and classifying them using convolutional neural network (CNN) and long short-term memory network (LSTM). To improve accuracy, model training is refined with the gray wolf algorithm. Evaluations with the InSDN dataset show this approach achieves ${9 9. 4 3 \%}$ accuracy with lower complexity and minimal time cost.