Adaptive Artificial Intelligence Techniques for DoS and DDoS Attack Detection in Modern Networks

The swift rise of connected systems, cloud technology, and widespread Internet of Things (IoT) infrastructure has led to an increased risk of campaigns of Denial of Service (DoS) and Distributed Denial of Service (DDoS) attacks. In fact, modern campaigns include elements of volumetric flooding, protocol abuse, application-layer attacks, and hacking of IoT devices while traditional methods based on rules and signatures have trouble with issues like high dimensional traffic and constant changes of behavior, slow speed of attacks, class imbalance, etc. In this review, the authors analyze the latest studies on DDoS detection methods referring to large databases, machine learning and deep learning systems, federated methods, and feature optimization strategies. The studies analyzed show considerable progress achieved after using novel systems, such as CNN, RNN, and others. The researchers note high effectiveness of using these approaches; however, there are still problems with efficient feature extraction and detection of minority classes. The authors further claim that efficient intrusion detection methods should be designed to be able to adapt to emerging threats.