Privacy-Preserving Popularity-Based Deduplication against Malicious Behaviors of the Cloud
Xiaowei Ge, Guanxiong Ha, Chunfu Jia, Zhen Su · 2024
Popularity-based secure deduplication scheme classifies data based on their number of owners and provides different levels of security for a trade-off between privacy preservation and storage savings. Most existing schemes rely on a trusted third party to record data popularity using deterministic tags, which is impractical in reality. Recently, Ha et al. propose a scheme that uses random tags to record popularity without the need for a trusted third party. However, their scheme is vulnerable to a malicious cloud launching smuggle attacks (SAs) and popularity-faking attacks (PFAs), which poses security vulnerabilities. In this paper, we propose a privacy-preserving popularity-based deduplication scheme. For one thing, we use unforgeable random tags to record data popularity, which defends against SAs. For another thing, we design a verifiable interactive popularity detection scheme to assure the correctness of popularity detection and resist PFAs. Security analysis and evaluation results show that our proposed scheme provides stronger security guarantees with limited overhead compared with existing schemes.