Implementation of Hyperparameter Tuning Random Forest Algorithm in Machine Learning for SDN Security: An Innovative Exploration of DDoS Attack Detection

Software Defined Network (SDN) represents a novel approach to network programming for the design, construction, and administration of computer networks. This approach entails the separation of the control plane from the data plane, with the objective of centralizing the network and consolidating all settings in the control plane. Network design frequently encounters challenges pertaining to network security and network attacks. SDN offers a number of advantages. One such advantage is that it allows the implementation of control functions at the controller level, thus positioning the controller as a key element of SDN. The centralized nature of SDN renders it susceptible to attacks, including Distributed Denial of Service (DDo <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$S$</tex>) attack. Consequently, a security system is required to address the issue of attacks directed towards the SDN controller. DDoS attack detection system is created using a machine learning algorithm, namely Random Forest, whose algorithm performance is optimized using Hyperparameter Tuning to achieve high accuracy in the detection process. The InSDN dataset, comprising a total of 56 classes, was employed in this research, and the SelectFromModel feature selection method was utilized to identify crucial features within the InSDN dataset. The method was found to be effective in detecting DDoS attacks, with an accuracy value of 99.99%. The results demonstrate that Hyperparameter Tuning Random Forest is an effective method for detecting DDoS attacks on SDN.