PriFR: Privacy-preserving Large-scale File Retrieval System via Blockchain for Encrypted Cloud Data

Abstract

As a fundamental and commonly used service, file retrieval has been extensively studied by information retrieval, cryptography, and big data communities. In this paper, we consider the problem of privacy-preserving file retrieval. A new framework named PriFR is proposed by integrating the blockchain and cloud computing infrastructures. The large-scale original files are encrypted and outsourced to the public cloud server. The encrypted retrieval indexes are stored on the full nodes in the blockchain to support traceable and unforgeable retrieval services. This design embraces the benefits brought by both cloud computing and blockchain. For the first time, PriFR decomposes the file retrieval problem into the numerical query and keyword search on the file metadata. In doing so, each file can be characterized more precisely than the traditional keyword search based schemes. In addition to functionality, PriFR only applies fast and lightweight symmetric cryptographical primitives to reach near plaintext retrieval efficiency. In specific, the numerical query is implemented atop order-preserving encryption (OPE) that is perfectly compatible with plaintext indexing techniques. The price for such high efficiency is its vulnerability to inference attacks. Given enough background knowledge, the plaintext can be recovered with high probability. To resist this attack, PriFR injects differential privacy noises into the raw data to offer guaranteed privacy-preserving strength with a negligible extra efficiency cost. The experimental results have demonstrated the effectiveness of PriFR.