Experimental analysis and comparative study of secure data outsourcing schemes in cloud

Abstract

When customers are transferring their private data to any third party, then there is much responsibility of both security and compliance. The outsourcing process is computationally secure if it is performed without unveiling to the other external agent or cloud, either the original data or the actual solution to the computations. Secure multiparty computation computes a certain function without revealing their private secret information. Homomorphic encryption is another solution that can deal with this situation. In homomorphic encryption, ciphertext (data in encrypted format) should be sent to the cloud, the computations are made on the ciphertext, and the result of this computation is a ciphertext form itself. If the result of the computation is decrypted, then the correct plaintext result must be obtained. In this paper, we summarised our proposed secure outsourcing computing protocols. As a cloud application, our proposed protocols encrypt data by different users and further are transformed to cloud. By utilising our protocols, we performed experimental analysis on virtual cloud to examine the novelty in terms of computational and communication complexity. Our implementations involve both CPU and GPU-based simulation results. We improved our adopted procedure to achieve better speed-up and security.