LAMET: A Defense-Centric Approach Against Economic Denial of Sustainability Attacks in Cloud Environments

Economic Denial of Sustainability (EDoS) attacks exploit the pay-as-you-go billing model in cloud computing to cause financial harm by artificially inflating resource usage and triggering unnecessary scaling. We propose a novel framework, the Learning Adaptive Mechanism for EDoS Threats (LAMET), which integrates advanced threat detection and adaptive security measures to mitigate such attacks. LAMET establishes Denial of Wallet attacks as precursors to EDoS, leveraging sustained financial anomalies that include repeated auto-scaling and resource overuse for early detection. Our approach combines anomaly detection using Isolation Forest with predictive behavioral analysis through Stochastic Recurrent Neural Networks, dynamically refining anomaly scores and adapting to evolving attack patterns. Extensive evaluations demonstrate LAMET’s superiority over state-of-the-art models, achieving a Mean Squared Error of 0.3164 and Root Mean Squared Error of 0.5625 while significantly reducing economic impact and false alarms.