An Innovative Hybrid System for Detecting Intrusions in Cloud Platforms

Abstract
Cloud computing's ongoing development continues to reveal major flaws in its architecture, emphasizing the need for stronger security measures. The complexity and rapid pace of modern cyber threats often render traditional intrusion detection systems (IDSs) ineffective. To enhance threat detection in cloud environments, we propose a hybrid intrusion detection system (IDS) framework that combines Hopfield Artificial Neural Networks (HANN) with Simulated Annealing (SA). Our approach provides excellent detection performance by integrating intelligent feature selection with HANN-based classification. When tested on the well-known KDD '99 benchmark dataset, the model achieved an exceptionally low false alarm rate of 0.36%, along with a detection rate of 99.9% and an accuracy of 95.2%. Notably, only 19 carefully selected features were required to achieve these results, demonstrating the framework's effectiveness. Compared to previous methods, this work marks a significant improvement and shows excellent promise for implementation in actual cloud security scenarios. Keywords - Intrusion Detection System, Cloud Security, Fuzzy Clustering, Hopfield Neural Network, Simulated Annealing, Feature Selection.