AI-Based Cloud-Edge Collaboration for Enhanced Cybersecurity in Industrial IoT

Abstract

The rapid proliferation of Industrial Internet of Things (IIoT) systems has posed significant cybersecurity challenges, as these devices are interconnected and distributed. Conventional security controls, which tend to be fixed and centralized, do not suffice to meet the dynamic, real-time needs of IIoT environments. The proposed paper presents a novel AI-assisted Cloud-Edge Collaboration Model to improve the cybersecurity of IIoT systems by leveraging edge computing to detect threats with low latency and cloud computing to scale and perform advanced analytics and threat intelligence aggregation. The proposed model is evaluated on the TON-IoT Network Intrusion Dataset, and the key performance metrics include Detection Accuracy, False Positive Rate (FPR), False Negative Rate (FNR), Response Time, and Scalability. The Proposed Model demonstrates significant improvements, with 98.6% detection accuracy, dramatically outperforming the Existing Model, with 90.5%. Also, the FPR decreases to 3.2 (instead of 7.5), and the FNR decreases to 2.4 (instead of 6.8), indicating improved performance in identifying true threats and reducing false alarms. The Proposed Model also shows that it has significantly shortened response time, which was 10-15 seconds in the Existing Model, to 3-5 seconds and is thus able to mitigate security threats faster. In addition to that, the number of devices supported by the Proposed Model is 500 while the Existing Model only supports 50 devices. These findings show the efficiency of combining AI, cloud, and edge computing in offering scalable, accurate, and efficient cybersecurity solutions for IIoT systems.