Enabling privacy-preserving multi-server collaborative search in smart healthcare

Abstract

With the advancement of smart healthcare, each medical institution stores huge amounts of users’ medical data in its cloud server for diagnosis and treatment. However, the traditional storage structure has obstacles to reasonable medical resources access and medical data circulation, such as data security, query privacy, and isolated data island problems. An intuitive solution is to store all the encrypted medical data in a central server. But this method absolutely depends on the central server, causing more performance disadvantages and privacy risks. Therefore, it is urgent to construct a secure and proper medical data retrieval scheme. Based on the existing data storage model, we build a multi-server search scheme to collaboratively perform diagnostic institution location, medical data search, and even cross-domain data search in this paper. The multi-server architecture solves problems of destructiveness and information over-centralization caused by the single server and enhances the reliability and practicality of the system. The utilization of hidden vector encryption, secret sharing, and secure multi-party computation realizes efficient search, identity privacy, search pattern security, and access pattern security. Security analysis demonstrates that identity privacy and query security are protected. Extensive experiments show that the scheme has better data search and data add efficiency through horizontal and vertical comparisons.