Attention Mechanisms with Deep Convolutional Neural Networks for Detection of Distributed Denial of Service Attacks
Asma Djama, Hamza Kheddar, Mohamed Maazouz · 2024
In the constantly evolving landscape of cybersecurity threats, distributed denial of service (DDoS) attacks continue to pose a significant challenge, disrupting services and causing substantial economic losses. Conventional detection techniques often struggle with the complexity and magnitude of DDoS attacks, which overwhelm target systems with excessive traffic. Existing methods frequently fail to detect these attacks accurately and promptly, leading to prolonged downtime. This paper proposes a novel solution by integrating Attention mechanisms with deep convolutional networks to enhance DDoS attack detection. The attention layers enable the framework to focus on the most pertinent features in the data, improving its ability to differentiate between normal and malicious traffic. We utilize the comprehensive CIC-IDS2017 dataset to assess our proposed approach. Key performance metrics, including accuracy, precision, recall, and F1 score, are employed to evaluate effectiveness. Our experimental results demonstrate that incorporating attention layers significantly enhances detection performance, achieving accuracy and precision rates of over 99.94%. This highlights the potential of our approach to provide a more effective and reliable solution for DDoS attack detection.