A Federated–Gan Hybrid Framework for Reducing Malware Attacks in Cloud Environments

Cloud computing has become a critical backbone for modern digital services, yet it remains highly vulnerable to increasingly sophisticated malware attacks. Traditional detection systems struggle with privacy concerns, limited training data, and the rapid evolution of malicious patterns. This study proposes a Federated–GAN Hybrid Framework designed to enhance malware detection and mitigation across distributed cloud environments. The framework integrates Generative Adversarial Networks (GANs) for synthetic malware generation and data augmentation with Federated Learning (FL) to enable decentralized, privacy-preserving model training. GANs strengthen the detection model by producing realistic malware variants that improve robustness against zero-day and metamorphic attacks, while FL ensures that sensitive organizational data remains local, thereby reducing privacy risks and communication overhead. The Design Science Research (DSR) methodology guides the development and evaluation of the hybrid model, ensuring a systematic approach to artifact construction and validation. Experimental evaluations demonstrate that the integrated framework achieves improved detection accuracy, faster model convergence, and enhanced resilience to adversarial threats compared to conventional centralized machine learning techniques. The proposed hybrid architecture provides a scalable, secure, and privacy-aware approach for reducing malware attacks in cloud environments, contributing both theoretical and practical advancements to cloud cybersecurity.