Sustainable Machine Learning for Real-Time DDoS Attack Detection and Mitigation in Industry 4.0 CPS

Industry 4.0, often known as the “Fourth Industrial Revolution,” has remodelled factories into “Sustainable Cyber-Physical Production Systems” (CPPSs) that connect workers, tools, and completed products. Benefits from digitization include introducing customer-centric, transparent, and adaptable production methods. However, it has also created new attack vectors, with malicious actors taking advantage of network security gaps and focusing on IoT devices. The intricate nature of Distributed Denial-of-Service (DDoS) assaults endangers production lines, corporate operations, and workers in today’s dynamic industrial context. To address these issues sustainably, this study employs Machine Learning (ML) to identify network anomalies and construct data-driven models for DDoS attack detection in Industry 4.0 CPPSs. Unlike previous methods, these use actual data collected from a semiconductor manufacturing facility’s network. During training, 45 aspects of network flow are extracted in both directions, and various labeled datasets are curated. Eleven distinct forms of monitoring strategies for supervised, unsupervised, and semi-supervised machine learning settings are examined in detail in this paper. The simulations’ accuracy indicates the efficacy of supervised algorithms, which are demonstrated to work with both loose and stringent supervision for detection. The Decision Tree model, for instance, achieves an outstanding $\mathbf{0 . 9 9 9}$ Accuracy and a truly remarkable $\mathbf{0 . 0 0 1 \%}$ False Positive Rate. In the face of increasing cyber threats, this long-term, sustainable strategy to cybersecurity in Industry 4.0 strengthens CPPSs’ resilience and security.