Detection of DDoS Attacks in Cloud Computing Environments Using Machine Learning Techniques

Cloud computing represents a paradigm shift in IT technology, providing end users with flexible, virtualized on-demand services that offer greater efficiency, reduced downtime, and lowered infrastructure costs. However, Distributed Denial of Service (DDoS) attacks remain a substantial threat to cloud reliability, causing severe damage and disruption. In DDoS attacks, perpetrators exploit compromised computers to inundate victim cloud infrastructures with packet streams, exploiting known or unknown vulnerabilities. These attacks can consume a substantial measure of the victim's network bandwidth utilization and server consumption time, significantly impacting performance. To mitigate the DDoS threat, this study proposes a detection method based on Random Forest Classifier and logistic regression algorithms. The Random Forest Classifier, when combined with detection algorithms, offers efficient DDoS attack detection. Evaluation results demonstrate substantial performance improvements in terms of precision, F1-score, sensitivity, and compared to several existing DDoS attack detection methods using real datasets.