Deep Robust Cramer Shoup Delay Optimized Fully Homomorphic For IIOT secured transmission in cloud computing

Several sensors obtain data during industrial outturn and send this collected data to the cloud server via Internet communication. Due to the reason that a cloud server is not an entirely trusted entity, data authenticity has to be provided prior to outsource to the cloud server, so that only authorized users or devices can access those authentic data from distinct topography areas. Hence, a strong privacy preservation mechanism is required during data collection. In addition, the data latency and network delay involved in communication should also be observed. In order to robust privacy preservation, Robust Cramer Shoup Delay Optimized Fully Homomorphic (RCS-DOFH) is proposed. This method includes three steps. First to minimize the communication overhead and time, Kullback–Leibler divergence is used in the Robust Cramer Shoup Decryption (RCSD) mechanism. Next, to minimize the data latency and network delay, Delay Optimized Fully Homomorphic Encryption (DOFHE) mechanism is designed. In this mechanism, delivery delay is calculated between the base station and IIoT device signal Finally, privacy preserving deep learning using RCSD and DOFHE is presented for privacy preserved secure data transmission. At first, RCSD mechanism is utilized to decrypt private generated signals along with its weight parameters. Then, the encryption is performed by using DOFHE mechanism. After that, activation function over the encryption region is determined. By this way, the proposed RCS-DOFH method achieves secure data transmission with minimum latency and network delay. The comparison of the RCS-DOFH method is provided and experiments conducted on SECOM dataset showed that the proposed method outperforms other conventional methods.