An enhanced deep learning integrated blockchain framework for securing industrial IoT

The Industrial Internet of Things (IIoT) is a collection of interconnected smart sensors and actuators with industrial software tools and applications. By gathering and evaluating industrial data in real time, IIoT seeks to improve manufacturing and industrial processes. However, a number of security risks might affect IIoT networks due to their heterogeneous and homogeneous characteristics. Data transmission over an unsecure communication channel leaves the possibility of malicious activity and communication between various entities being accessed by hackers. Thus, protecting the privacy and security of data transferred via IIoT networks is crucial. Motivated by the aforementioned challenges, this paper introduces a deep-learning-based blockchain architecture for IIoT network security. Initially, the industrial devices are registered in the IIoT, and after that, a unique identification (ID) is created for each device presented in the IIoT. For that, the security algorithm named Secure Hash Algorithm-256 (SHA-256) is used to generate unique keys for each device. After that, the Deep Neural Network (DNN) algorithm is developed to perform the encryption and decryption processes for each data record of the IIoT device. After encryption, the secured data of each IIoT device is stored in different blocks for further data communication. Finally, Enhanced Bidirectional Long-Term Memory (EBLSTM) is developed to validate whether the data presented inside the block is valid or not. The TON_IoT dataset is utilized in this research to perform the validation process, and the performance is evaluated through different performance parameters. The results show that the proposed method accomplishes 123tps throughput for a 4.5MB block size with 800 transactions and accomplishes 0.4% authentication accuracy.