Secure search for encrypted personal health records from big data NoSQL databases in cloud

Abstract

As the healthcare industry adopts the use of cloud to store personal health record (PHR), there is a need to ensure that we maintain the ability to perform efficient search on encrypted data (stored in the cloud). In this paper, we propose a secure searchable encryption scheme, which is designed to search on encrypted personal health records from a NoSQL database in semi-trusted cloud servers. The proposed scheme supports almost all query operations available in plaintext database environments, especially multi-dimensional, multi-keyword searches with range query. Specifically, in the proposed scheme, an Adelson-Velsky Landis (AVL) tree is utilized to construct the index, and an order-revealing encryption (ORE) algorithm is used to encrypt the AVL tree and realize range query. As document-based databases are probably the most popular NoSQL database, due to their flexibility, high efficiency, and ease of use, MongoDB, a document-based NoSQL database, is chosen to store the encrypted PHR data in our scheme. Experimental results show that the scheme can achieve secure and practical searchable encryption for PHRs. A comparison of the range query demonstrates that the time overhead of our ORE-based scheme is 25.5% shorter than that of the mOPE-based Arx (an encrypted database system) scheme.