Harris Corner Detection Algorithm based Long Short-Term Memory for Detecting Distributed Denial of Service Anomalies in Software Defined Networks

Software-Defined Networking (SDN) is the innovative development in network technology with a number of attractive features, including management and flexibility. Despite these advantages, SDN was compromised by Distributed Denial of Service (DDoS) attacks, which causes substantial challenge due to the harm networks. Hence, innovative model is required to identify DDoS attacks despite a range of security methods. In order to detect DDoS attacks in SDN in real time, the Harris Corner Detection Algorithm-Long Short-Term Memory (HCDA-LSTM) is presented in this research. The Canadian Institute of Cybersecurity Intrusion Detection system (CICIDS) 2017 dataset is initially used to collect the data. Consequently, the proposed HCDA-LSTM model significantly reduced the challenges of network security, particularly in relation to SDN. The proposed method enhances SDN security by mitigating DoS/DDoS attacks and improve the performance effectively in results. When compared to the existing models such as Deep Learning Approach for Detecting Controller (DLADSC), Extreme Gradient Boosting (XGBoost) and XRDI the suggested HCDA-LSTM method achieved better performances in terms Accuracy in CICIDS 2017 dataset of 99.92%, Precision of 99.64%, F-measure of 98.89% and recall of 98.65% which is comparatively higher when compared to existing models.