Research on Detecting DDoS Attacks on SDN Platform Based on Bi LSTM Algorithm
Sui Xiangning, Qinan Li · 2023
Low-rate Distributed Denial of Service ($L$DDoS) attack has significant features of periodic attack traffic due to its small attack traffic, strong concealment, and huge harm. There is a deficiency in L DDoS attack detection in the existing SDNs that ignores temporal feature detection in attack traffic. While the Bi LSTM algorithm has advantages in temporal feature detection. Based on it, L DDoS attack detection scheme in SDN platform is proposed. Firstly, the five feature vectors in the flow table of the SDN switch-timestamp, average number of flow packets, average bit of flow packets, port rate, flow rate and source rate are calculated in real-time computing. Then they are sent into the Bi LSTM model to judge abnormal traffic and temporal feature detection for predicting the periodic time of L DDoS. Hence, it can make accurate detection of L DDoS, improving the accuracy and real time. The results show that the accuracy of L DDoS abnormal traffic detection reaches 99.2%.