With the explosive growth of data, it has become increasingly popular to deploy the powerful cloud to manage data. Meanwhile, as the cloud is not always fully trusted, personal and sensitive data have to be encrypted before being outsourced to the cloud. Naturally, this brings a serious challenge for the cloud to provide secure and efficient query services over huge volumes of data. Although existing works have proposed some solutions to solve the above challenge, most of them just focus on the single keyword query and cannot directly support multi-keyword query. Even though some works have discussed solutions for the multi-keyword query, they cannot well balance the efficiency and privacy. Therefore, in this paper, we propose a novel multi-keyword conjunctive query scheme over cloud, which can achieve high query efficiency with small privacy leakage. In specific, we first design a tree-based index to support the multikeyword conjunctive query and employ Boneh-Goh-Nissim (BGN) homomorphic encryption technique to protect its privacy. Then, based on the tree-based index, we propose a wildcard search algorithm to improve its query efficiency. Finally, the detailed security analysis shows that the proposed scheme is really privacypreserving, and extensive simulation results also demonstrate its efficient.

Introduction

With the rapid development of the internet, volumes of data are exploding by the day. According to IBM Marketing Cloud study [1], more than 90% of data on the internet has been created since 2016, which leads more and more individuals and companies to store local files in the cloud to reduce the increasing storage overhead. However, the cloud servers may not be fully trusted in practice today, because administrators or even hackers are likely to get full access to the servers and consequently to the files. Thus, the files with some sensitive information (e.g., electronic health records) have to be encrypted before outsourcing them to the cloud. Although data encryption technique preserves data privacy, it also hides some critical information such that the cloud cannot support some user’s operations over the encrypted data, e.g., as multi-keyword conjunctive query, which returns a set of files containing multiple queried keywords. Consequently, it is challenging to perform multi-keyword conjunctive query over encrypted data. One straightforward solution is that the user downloads all encrypted files and performs search after files are decrypted. However, this solution is impractical because of its significant computational cost and communication overhead. In order to solve the above problem, searchable encryption (SE) was introduced, which allows the cloud server to search encrypted files without leaking information in the plaintext files.