Modified DNA-based Cryptography System in the Cloud: Deep Maxout-based Fined Tuned Key Generation

Abstract

Cloud security is a set of practices and tools created to address both internal and external security threats to businesses. Organizations must have cloud security as they implement their digital transformation schemes and include cloud-based tools and services in their infrastructure. Cryptography is a mechanism for preventing illegal access to data. In this paper, modified DNA-based cryptography (MDNAC), which is defined as data hiding with respect to DNA sequence is used. The steps involved in the proposed MDNAC is: encryption and decryption with optimal key generation. A way of converting plain text into cipher text is known as encryption. Two components make up the encryption process: a key and an encryption algorithm. For the encryption algorithm, we employed a modified DNA algorithm. In the decryption phase, the reverse operation is performed to get the plain text from the cipher text. Moreover, a deep learning model is used for generating the keys; the model used is deep maxout. To ensure the appropriate key generation process, the weights of the deep maxout are optimally tuned by the new feedback assisted Archimedes optimization (FAAO) algorithm. Based on the generated keys, the encryption process takes place. Finally, the performance of MDNAC is evaluated using conventional methods with respect to different measures. Additionally, the MDNAC obtained a correlation value of 0.20297 for the mean case scenario, despite the fact that the corresponding values are 0.02%, 0.17%, 0.2%, 0.7%, 0.12%,0.41%, 0.86%, and 0.46% as compared to the other models such as FAT, AOA, BMO, COOT, BOA, SSO, WOA, and LES respectively.