Privacy-preserving verifiable fuzzy phrase search over cloud-based data

Abstract

Today, cloud computing plays an increasingly important role in cloud-based data processing systems. Privacy-preserving phrase search is one critical search technique in cloud-based data processing systems, which allows for the retrieval of cipher documents containing a set of consecutive keywords. However, the existing privacy-preserving phrase search schemes rarely support fault tolerance and result verification at the same time. To deal with these problems, this paper presents a Privacy-Preserving Verifiable Fuzzy Phrase Search scheme over cloud-based data (PPVFPS). We construct a novel keyword matching list by the keyword transformation techniques and the secure $kNN$ algorithm to support fuzzy search. To enhance search efficiency and achieve the dynamic update, we generate a counting Bloom filter based on the virtual binary tree, which can help find the documents containing the search keywords. In order to securely judge the position relationship between two search keywords, we employ the techniques of homomorphic encryption and bilinear mapping to encrypt the positions of the keywords. We generate the verification tags based on the MAC technique and build a document index table to realize result verification. We demonstrate the security analysis of PPVFPS scheme, and the experimental result shows that PPVFPS scheme can achieve high accuracy.