A Robust Lattice-Based Protocol for Outsourcing Scientific Computations in Multi-Cloud Architecture

Abstract

The research delivers an enhanced lattice-based encryption protocol which operates specifically for scientific computing processing in multi-cloud systems. The system employs two independent cloud servers to provide secure data processing with privacy features and efficient operation. The users employ a probabilistic lattice-based encryption method to secure their private data while the scheme protects against quantum attacks and provides semantic security. The encryption method uses homomorphic transformations to process input data as encrypted messages that protect the original text alongside the decryption keys. The architecture supports parallel processing and dynamic scaling, enabling secure multi-user collaboration across heterogeneous computational environments. Performance evaluation of the protocol shows how it functions under different data size conditions from kilobytes to gigabytes and cryptographic key length settings from 16 to 256 bits. The proposed solution outperforms customary homomorphic systems by reaching a 35-40% reduction in operational cost with high-throughput scenarios resulting in a 50% shorter upload time and decryption duration. Both multilevel encryption and the decryption procedure operate in phases which permits authorized users to recover final data while blocking unauthorized access for confidentiality and proper access control. The model enables substantial enhancement of reliability and fault tolerance and security for real-time applications running across multiple cloud infrastructures.