A hybrid wrapper technique enabled Network Intrusion Detection System for Software defined networking based IoT networks

In the age of smart era, usage of IoT devices is inevitable. As the number of IoT devices increases, the amount of data they generate also increases, which in turn leads to security breaches. Continuous monitoring through SDN is the appropriate solution to mitigate security breaches. SDN introduces a new architecture where the control plane operates independently of the data plane. While this design allows applications in the base plane to interact with network components via the controller, it also creates security concerns. An attacker could exploit vulnerabilities in the control plane, potentially leading to significant data breaches. Amid a lot of research in various venues to address the security aspects of SDN, a simple and robust Intrusion detection system is necessary for the current context. Hence, a hybrid-wrapped enabled feature selection RF-RFE has been included in determining the highly influential features upon the derived features and a fine-tuned machine learning-based model to detect the anomalies at the earliest possible in the context of SDN has been trained and tested on SDN-IoT dataset. The tested results reveal that RF-based RF-RFE-SDNIoT-NIDS obtained 99% of its binary and multi-classification accuracy using minimal features compared to other machine learning models.