Optimized Hybrid Multi Component Attention Graph Convolutional Neural Network and Multikey Verifiable Homomorphic Encryption based Complex Cloud Systems

The numerous user transactions that occur within the system, data security is a primary concern in cloud computing. Cryptographic analysis and advancements are important areas of current research. To overcome these issues, a number of security mechanisms have been implemented place, like knowledge-based authentication, two-factor authentication, adaptive authentication, multi-factor authentication, and single-password authentication. This paper proposes an Optimized Hybrid Multi Component Attention Graph Convolutional Neural Network and Multikey Verifiable Homomorphic Encryption based Complex Cloud Systems (HMCAGCNN-MVHE-CCS-RBMO). Here, the input data is gathered from cloud storage and is divided into three categories: cloud storage level, security level, and user level. Key generation and encryption operations are carried out at the user level. Consequently, key-level management is used to control the degree of security for the data uploaded to the cloud in encrypted form, utilizing Hybrid Multi-Component Attention Graph Convolutional Neural Network and Multi-key Verifiable Homomorphic Encryption (HMCAGCNN-MVHE) for the encryption process. Therefore, it is proposed to utilize the Red-billed Blue Magpie Optimization Algorithm (RBBMOA) enhancement the HMCAGCNN-MVHE for cloud complex systems. The proposed method, HMCAGCNN-MVHE-CCS-RBMO is implemented and its efficiency is evaluated based on several performance metrics, like encryption time (ET), cipher text, throughput, bit error rate (BER) and computation time. Then the performance of the proposed HMCAGCNN-MVHE-CCS-RBMO method is linked by the existing methods: privacy-preserving deep neural network processing for credit risk prediction in cloud: a homomorphic encryption-dependent method (PDNN-CRP-HEN), Secure Dynamic Encrypted Property Graph Search With Lightweight Access Control (SDEG-LAC) and decentralized federated graph learning along lightweight zero trust architecture for nextgeneration networking security (DFGL-ZTA-GNS) respectively.