Monarch butterfly optimisation algorithm-based Twofish encryption algorithm for secure data storage management in cloud computing and intrusion detection through deep learning

Cloud computing techniques face issues while storing large amounts of data. To overcome this, the deep learning-based Recurrent Neural Network (RNN) with a Monarch Butterfly Optimisation Algorithm (MBOA)-based Twofish encryption algorithm is developed. Initially, user data is split into packets and transferred over the Internet. The data packets are pre-processed to handle redundant data and remove null values. After preprocessing, the appropriate characteristics are selected using Pearson’s correlation. An RNN is applied to the extracted features to determine whether the user data is attacked or normal. Since the packet data is still quite large, B-Tree Huffman Encoding (BHE) is used to compress the packet data for better compression. Finally, the compressed data is encrypted using the MBOA-based Twofish algorithm for secure storage in the cloud. In this paper, MATLAB software was used to implement and evaluate the performance. From the experimental analysis, the proposed model achieved 94.4% accuracy, 96% precision, 93% sensitivity, 96% specificity, 95% F1-score, and 89% kappa, along with a low false positive rate of 4.1% and an error rate of 5.6%. Furthermore, the optimised encryption achieves 46 seconds of encryption time, 16 seconds of decryption time, 61 seconds of execution time, and only 2% information loss.