An Efficient Graph-Based Approach for Identifying Attack Patterns in IoT and IIoT Networks

The growing attack surface and interconnection of the Internet of Things (IoT) and Industrial IoT (IIoT) have presented major cybersecurity issues. This work presents a graph-based analytical approach for IoT and IIoT environment network vulnerabilities assessment and attack pattern recognition. Synthetic graphs were created with the Edge-IIoTset dataset to simulate network traffic in both normal and attack conditions, encompassing Distributed Denial of Service (DDoS) and backdoor assaults. The framework utilizes centrality measurements and temporal traffic analysis to identify crucial nodes, access points, and compro-mised devices. This method offers practical insights for intrusion detection and threat mitigation, emphasizing the capability of graph-based techniques to improve 1oT/IIoT network security. The findings illustrate the scalability and feasibility of the framework for real-time security applications in industrial environments.