Enhancing Cloud Security with Intrusion Detection and Prevention Systems (IDPS): Comparative Evaluation of Signature-Based, Anomaly-Based, and AI-Powered Detection Models

Abstract

The rapid adoption of cloud computing has amplified cybersecurity threats, necessitating robust Intrusion Detection and Prevention Systems (IDPS). This study aims to comparatively evaluate signature-based, anomaly-based, and AI-powered detection models in enhancing cloud security. Employing a mixed-methods approach, we utilized real-world datasets such as NSL-KDD and CIC-IDS2017 to simulate cloud environments. Signature-based models excelled in detecting known attacks with 95% accuracy but faltered on zero-day threats. Anomaly-based systems identified novel intrusions at 85% precision, though prone to false positives. AI-powered models, leveraging machine learning algorithms like Random Forest and LSTM, achieved superior performance with 98% detection rates and reduced false alarms. Findings underscore AI's transformative potential in adaptive threat mitigation. Key conclusions highlight the need for hybrid IDPS frameworks to balance speed, accuracy, and scalability in cloud ecosystems, informing policy and practice for resilient infrastructure.