Real-Time Malicious Intrusion and Attack Detection in IoT-Enabled Cybersecurity Infrastructures

Internet of Things (IoT) networks present's unique challenges in cybersecurity due to their distributed and dynamic nature, making them highly vulnerable to botnet attacks. Existing defense mechanisms often struggle to accurately distinguish between benign and malicious traffic, leading to suboptimal detection accuracy and high false alarm rates. To address this, we propose the Botnet Attack Detection and Defense (BADD) mechanism, a supervised learning-based approach utilizing Convolutional Neural Network (CNN) models. BADD operates by extracting parametric features from traffic data buffered within fixed time frames, enabling predictive analysis to identify potential botnet attacks. We experimented on benchmark datasets with four different CNN models and got encouraging results. The trained models exhibited training accuracies ranging from 0.852 to 0.857 and testing accuracies between 0.825 and 0.862. The effectiveness of our method for detecting harmful intrusions in real-time in cybersecurity infrastructures enabled by the Internet of Things is demonstrated by a comparative analysis with modern methodologies.