An Efficient DDoS Attack Detection Mechanism in SDN Environment
As cyberattacks get more sophisticated, it becomes more difficult to identify advanced attacks in a wide variety of industries, including business, national defense, and healthcare. Traditional intrusion detection systems are insufficient to identify these sophisticated attempts with unpredictable patterns. Attackers evade recognized signatures and spoof legitimate users. While Software-Defined Network (SDN) provides greater innovation to the design of future networks, it is also more susceptible to network attacks. We present a system for the detection and protection of network attacks in the SDN environment using Deep Learning (DL) for addressing challenges. We use InSDN, the most recent IDS dataset specially developed for the SDN environment. InSDN includes sophisticated DoS and network attacks and, it is a publicly accessible dataset. Additionally, we compare the model performance trained using CIC-IDS2017 and CIC-DDoS2019 datasets respectively. We focus on the DDoS attack category in this study and create an intrusion detection DL model for SDN. We construct our model using a Convolutional Neural Network (CNN) and assess its performance. Additionally, we propose the optimized CNN design for improved performance based on various evaluations.