SecureIIoT Environment: Federated Learning Empowered Approach for Securing IIoT from Data Breach

The growing congruence of gadgets today resulted in a numerous type of cyber attacks. A similar trend occurs with the industrial Internet of Things (IIoT), wherein increasing data created by connected equipment offers up new opportunities for enhancing service quality for new applications. But, security has become a major design priority for smart devices since the introduction of the IIoT. However, data providers have significant challenges in exchanging their data between different cyber physical systems due to privacy and security concerns (e.g., data imbalance and leakage). In this article, we developed a safe data sharing architecture for various IIoT devices using federated learning (FL). The proposed architecture incorporates FL into the edge computing consensus process, allowing the consensus computing activity to be used for federated training as well. The proposed framework achieves high efficiency, and better security, according to numerical findings generated by experimenting deep learning models. More precisely, the proposed framework named as SecureIIoT, is able to achieve 99.79% accuracy by detecting attacks as a binary classification problem.