Industrial Wireless Internet Zero Trust Model: Zero Trust Meets Dynamic Federated Learning with Blockchain
Haoran Xie, Yujue Wang, Yong Ding, Changsong Yang, Hai Liang, Bo Qin
Abstract
As a critical infrastructure for contemporary information technology industry, industrial internet of things (IIoT) contains a vast amount of sensitive data, making it a key requirement to ensure data security. As the use of wireless networks as a means of communication between nodes is becoming more and more common, in order to prevent malicious attacks from compromising the system, a zero-trust authentication system is necessary. In this article, we propose a comprehensive implementation framework for zero-trust verification of IIoT wireless transmission nodes, which utilizes federated learning to achieve zero-trust rule training and terminal model training, while employing blockchain technology for on-chain aggregation and cloud backup of the models. This approach enhances the accuracy and availability of the zero-trust rules while safeguarding the security of IIoT nodes. The constructed zero-trust framework incorporates a self-incremental learning function, and experiments show that it achieves a high level of accuracy at recognising attacks. Finally, we discuss the challenges of utilizing federated learning in zero-trust for IIoT and several potential solutions to address these challenges.