AN IMPROVED ECC ALGORITHM FOR SECURE CLOUD STORAGE SYSTEM WITH THE HELP OF SHA-256 BASED USER AUTHENTICATION AND DEEP LEARNING BASED INTRUSION DETECTION SYSTEM
A cloud environment is a collection of resources used to provide on-demand services to cloud clients. Access to the cloud environment is supplied through internet services, and data kept in the cloud environment is more accessible to both internal and external invaders. At the moment, cloud security experts are unable to provide a more dependable, safe, and effective intrusion detection system (IDS) for identifying intruders in data transfer. For fulfilling this requirement, in this paper a novel deep learning (DL) model namely Wavelet Kernel including Multi-Layer Perceptron Neural Network (WKMLPNN) based IDS is proposed for identifying the attacks while performing cloud data transmission. Additionally, to offer cloud secure storage, Improved Elliptical Curve Cryptography (IECC) is proposed with a novel key generation mechanism. Initially, the cloud user or data owner (DO) is registered with the Cloud Service Provider (CSP) for uploading their data to the cloud. Then at the CSP end, the authenticity of the DO is checked by implementing the Secure Hash Algorithm 256 (SHA-256) technique. After successful authentication, the DO is allowed to upload their data to the cloud. The CSP receives the data from the DO and then the intrusion present in the received data is identified using novel IDS (WKMLPNN). The presented IDS follows three operations such as preprocessing, optimal feature selection using Time factor incorporated Gravitational Search Algorithm (TGSA) and WKMLPNN applied classification for identifying the intrusions present in the data. If the received data contain the intrusion, the data will be avoided and not stored in the cloud, otherwise, the data will be stored in the cloud securely using the IECC algorithm. Finally, if the cloud users request CSP for data retrieval, the authenticity of the user is again checked using the SHA-256 algorithm. If the user is an authenticated one, the retrieval permission is granted for him, otherwise, the permission is denied for the user. The simulation results reveal that the presented approach outperforms existing strategies for intrusion categorization and encryption.
Cloud Computing, Cloud Security, Intrusion Detection System, Encryption Algorithms, Deep Learning, User Authentication.