Robustness Enhanced Proactive Fault-Tolerant Framework in Industrial Edge Networks

With the development of the Industrial Internet of Things (IIoT), proactive fault tolerance through multinode collaboration has emerged as a key approach to ensuring system stability. However, the distributed nature of edge environments introduces significant challenges to the robustness of existing proactive fault-tolerant systems. Outside the system, malicious nodes may disrupt the fault tolerance process, necessitating a robust collaborative mechanism to mitigate their impact. Inside the system, frequent node failures and other dynamic factors result in a highly dynamic network topology, requiring robust methods to optimize the effectiveness of fault identification and task migration decisions. In this article, we utilize blockchain and generative adversarial network (GAN) to construct a robustness enhanced proactive fault-tolerant framework. In our framework, we use blockchain for edge node supervision, and design an on-chain state lock mechanism to ensure the reliability of task migration during fault-tolerance processes. Considering QoS objectives and the credibility evaluations of blockchain on edge nodes, we construct proactive fault-tolerant task migration problem formulas and design a robust GAN-assisted proactive fault-tolerant task migration decision method based on these formulas. Finally, in an edge network built with Raspberry Pi devices, we validated the robustness of the proposed framework and the effectiveness of the proposed scheduling method. Compared with the baseline method, our method improved the task completion rate by an average of 13.8%, and reduced task completion delay and energy consumption by an average of 24.5% and 6.8%, respectively.