A secure cloud data deduplication method based on SGX under cloud-edge collaboration

Abstract

The continuous growth of data places increasing demands on cloud storage space, security, and performance. Safe deduplication technology can effectively reduce storage space requirements while ensuring the security of data deduplication for cloud computing. However, the current mainstream cloud security deduplication solutions rely on trusted third parties to ensure key security, introducing additional trust assumptions and key management overhead. And the deduplication schemes of the two parties are facing challenges in terms of security and transmission costs. In response to this issue, this article proposes a software protection extension based on the Intel software guard extension (Intel SGX) cloud data security deduplication method, the main innovation of which lies in: (1) designing an edge-based cloud edge collaborative security architecture for deduplication which reduces transmission overhead between cloud and edges, and the use of SGX trusted execution environment ensures the security of data and keys in cloud and edge. This does not require an untrusted third-party key server. (2) Propose a low-cost key security generation protocol, which generates keys through cloud edge collaboration, generating unique encryption keys for each edge user instead of individual keys for each data. (3) Design an efficient tag query mechanism based on tag popularity. By calculating the tag frequency, the separation of hot and cold tags can be achieved. This can improve tag inquiry efficiency while avoiding label misidentification. Theoretical proof of security and system experimental testing indicate that the proposed solution in this paper is superior to other security deduplication methods. It greatly reduces the cost of key management and label query without compromising security, among which the cost of key storage can be saved 48.72%∼98.72%.