Enhancing Intrusion Detection and Cloud Security by Integrating Snort with Advanced AI Techniques for Improved Accuracy and Threat Mitigation

Abstract

Intrusion Detection Systems (IDS) are critical for ensuring the security of cloud infrastructures and modern networks, which are increasingly vulnerable to sophisticated cyber threats. While traditional IDS like Snort effectively detect signature-based attacks, they struggle with high false-positive rates and limited adaptability to evolving threats. To address these challenges, this research introduces AIML-Snort, a hybrid intrusion detection framework that integrates Snort with advanced Artificial Intelligence (AI) techniques to enhance detection accuracy and mitigate false alarms. The framework employs preprocessing techniques such as feature normalization, outlier detection using Isolation Forest, and Recursive Feature Elimination (RFE) combined with Genetic Algorithms (GA) for optimized feature selection. Machine learning models like Random Forest (RF) and Neural Networks (NN) are then trained to identify anomalies and integrated with Snort’s rule-based system to analyze unresolved traffic. Using the UNSW-NB15 dataset, the framework was evaluated in a simulated high-speed network environment designed to emulate real-world traffic conditions. Results demonstrate that AIML-Snort outperforms traditional Snort configurations, achieving higher detection accuracy, significantly reduced false-positive rates, and improved scalability in large-scale deployments. The integration of AI techniques enhances Snort’s ability to adapt to dynamic and previously unseen threats, making it a robust solution for cloud security. This research establishes AIML-Snort as a practical and scalable IDS framework for addressing modern network security challenges, providing valuable insights into the synergy between traditional rule-based systems and AI-powered approaches. Future work will focus on further optimization and extending the framework to support additional datasets and attack scenarios.