Key points are not available for this paper at this time.
ABSTRACT The growth of the Internet‐of‐Things (IoT) and edge computing has led to new vulnerabilities and higher risk through the potential for distributed denial of service (DDoS) attack vectors against hosting IoT devices and edge nodes. Because of the significant burden imposed on resources, countermeasures for DDoS attacks do not perform well in resource‐constrained IoT and edge computing. This systematic review examines state‐of‐the‐art Machine Learning (ML) and Deep Learning (DL) techniques for DDoS attack detection and mitigation in IoT and edge systems, focusing on advancements from 2020 to 2025. We propose a novel classification framework based on deployment architecture, learning paradigm, attack specificity, and performance considerations. Our analyses encompass 33 high‐quality studies, revealing that 60.6% employ DL methods, 30.3% leverage ML techniques, and the remainder combine both, highlighting the superiority of DL models in detection accuracy notwithstanding their computational overhead, with ML methods offering lightweight alternatives and trade‐offs in adaptability. Furthermore, we evaluate public datasets and simulation tools used for validation. This review covers critical issues such as inference latency, energy consumption, and heterogeneity of deployment, together with solutions to some problems emerging, including tiny machine learning and federated learning. Finally, we identify gaps in the literature, accentuating the need for up‐to‐date datasets, adversarial resiliency, and detection‐mitigation frameworks. This review provides a comprehensive resource for engineers, researchers, and practitioners in the domain of DDoS defense in IoT and edge computing environments.
Opoku et al. (Mon,) studied this question.