Real-Time Hybrid Machine Learning-Based Next-Generation Intrusion Detection System for Edge Computing Networks

Abstract

With the rapid advancement of network technology, attacks such as denial-of-service (DoS), distributed denial-of-service (DDoS), and unknown or emerging threats are becoming more complex and harder to detect using conventional methods. Traditional intrusion detection systems (IDS) often struggle to detect attacks in high-speed real-time networks due to their reliance on rule-based or signature-based detection methods and limited processing speed. Additionally, organizations and industries are facing significant challenges due to the expansion of electronic devices like the Internet of Things (IoT) and running computer applications. There is an urgent need to secure these sensitive IoT high-speed network traffic systems. To address these limitations, this study develops a reliable and effective Hybrid Machine Learning-based Real-time Intrusion Detection System (HMLRT-IDS) to detect DoS, DDoS, and emerging attacks in real-time network traffic. The study proposes a Python programming-based algorithm for feature extraction that sniffs data from real-time network traffic and extracts relevant data related to the features, enabling the successful identification of cyber threats in real-time network traffic. Moreover, the Real-time Network Intrusion Detection-23 (RTNID23) dataset is constructed to evaluate the proposed HMLRT-IDS. Experimental results demonstrate that HMLRT-IDS achieves a remarkable accuracy of 99.88% with a response time as low as 0.001 seconds, making it highly efficient for real-time applications. Additionally, the proposed system outperforms several existing methods in terms of accuracy, speed, and adaptability, offering a significant advancement in intrusion detection for high-speed IoT or edge computing networks. This model simplifies the analysis of attacks and aids in determining the appropriate response when an attack occurs.