Detecting and Identifying the Targets of Covert DDoS Attacks

Network systems are essential to our daily lives but remain vulnerable to Distributed Denial-of-Service (DDoS) attacks, particularly stealthy low-rate variants that evade conventional detection methods. This paper presents an SDN-based deep learning framework designed to detect adaptive low-rate DDoS attacks targeting both end hosts and network links. Our approach not only mitigates these threats but also differentiates between host-targeted and link-targeted attacks, effectively countering dynamic adversaries. We construct dataset of benign and low-rate DDoS traffic and evaluate our solution in an SDN environment using the Mininet emulator and RYU controller, demonstrating its efficacy in identifying and countering sophisticated low-rate attacks.