DESIGN OF A ROBUST DEEP LEARNING FRAMEWORK FOR REALTIME CYBERATTACK DETECTION IN CLOUD COMPUTING

Abstract

Cloud computing has become a backbone for modern digital services, but its distributed and virtualized nature makes it highly vulnerable to cyberattacks. Traditional security mechanisms often fail to detect sophisticated and zero-day attacks in real time. This paper proposes a robust deep learning–driven framework for real-time cyberattack detection in cloud computing environments. The framework integrates automated data preprocessing, feature optimization, and a hybrid deep learning architecture to accurately identify malicious activities. Various attack categories such as denial-of-service, probing, user-to-root, and remote-to-local attacks are analyzed. Experimental results demonstrate improved detection accuracy, reduced false alarm rates, and faster response time compared to conventional machine learning methods. The proposed model enhances cloud security while maintaining scalability and efficiency. The framework is suitable for real-world cloud infrastructures requiring continuous monitoring.