Modeling and analysis of cascade failures in Industrial Internet of Things based on task decomposition and service communities

In Industrial Internet of Things (IIoT) systems, the overload of certain nodes may lead to the failure of the entire network, a phenomenon known as cascade failures. This occurrence has an immeasurable impact on industrial production. Establishing a realistic IIoT cascade failure model is crucial for enhancing the cascade reliability of IIoT systems. However, existing research on cascade failures in IIoT lacks an in-depth exploration of the actual characteristics in industrial scenarios, hindering accurate modeling of the cascade failure process in IIoT. To better characterize the cascade failure process of IIoT, this study proposes an interdependent network model from a cyber-service coupling perspective, considering task decomposition, service community structure, and coupling patterns. On this basis, we design a realistic cascade failure model that is based on production supply relationships between manufacturing units. In the experiments, we first validate the rationality of the proposed model through load distribution and load probability density analysis. Furthermore, the study of key modeling parameters then reveals that cyber network attacks cause more damage than service network attacks. Finally, we apply the model to four industrial manufacturing scenarios and explore their cascade failure performance.