Privacy-Preserving Data Synchronization Using Tensor-Based Fully Homomorphic Encryption

Abstract

A connected environment brings a challenge of the privacy for data synchronizations as multiple unexpected/unpredictable parties maybe involved the process of the data usage. Outsourcing tasks results in data uncontrollability and privacy concerns, since service vendors generally have an access to the data stored in the server. Thus, unencrypted data in the third-party server are facing a threat of the privacy leakage due to various potential causes. In this paper, we focus on the privacy leakage issue in data synchronizations, more specifically speaking, to address designing a privacy-preserving method for data multi-storage with a homomorphism capability. The proposed scheme utilizes a tensor-based Fully Homomorphic Encryption (FHE) that balances privacy protections and functionalities. Design objectives are threefold: the proposed method can (i) protect data owners' privacy, (ii) support arithmetic operations, and (iii) achieve real-time data synchronization. The expected application scenario is synchronizing owner's data with remote servers in cloud computing. Both theoretical proofs and experiment evaluations have been processed in order to examine the adoptability and correctness of our approach.