Deep Defender: An Adaptive Edge Cloud Framework for Intelligent DDoS Detection, Deception, and Tracking

Abstract

Cloud computing environments are increasingly vulnerable to Distributed Denial-of-Service (DDoS) attacks and sophisticated cyber intrusions that compromise service availability, scalability, and operational security. In this research, we have developed a resilient cloud-based intrusion detection framework using a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) deep-learning architecture for intelligent DDoS detection and cyberattack tracking. The proposed framework integrates spatial feature extraction and temporal traffic-learning mechanisms to effectively capture complex network-behavior patterns within heterogeneous cloud environments. We used the NSL-KDD dataset for experimental validation, which included preprocessing, feature normalization, correlation analysis, feature-importance evaluation, PCA visualization, and t-SNE-based nonlinear traffic analysis. The results reflected high detection performance (accuracy-99.14%, precision-99.06%, recall-99.14%, and F1 score-99.04%) besides low false-positive behavior and strong attack-prediction. The developed framework exhibited robust adaptive intrusion-learning capability, reliable multi-class attack classification, and scalable operational suitability for intelligent cloud-security monitoring and resilient cyber-threat management applications.