DMASP: Dynamic Multi-keyword Searchable Encryption for Protected Access and Search Patterns with Differential Privacy

In recent years, cloud computing services have grown rapidly, with people outsourcing huge amounts of private data to cloud servers. Searchable encryption(SE) facilitates people’s use of data while protecting data privacy. To balance the efficiency, current SE schemes still leak information such as access, search and volume patterns to cloud servers, and most dynamic SE schemes leak forward and backward privacy, and these leaks create serious security threats. In this paper, we propose a dynamic SE scheme DMASP with suppressed leakage, which protects access pattern, search pattern and volume pattern simultaneously. We utilize the differential privacy mechanism to obfuscate databases, and introduce the CKKS algorithm to protect access and volume patterns during queries. Meanwhile, we design a secure structure to index databases efficiently, and protect forward and backward privacy during updates. We rigorously analyse the security of DMASP. Furthermore, comprehensive experimental evaluation show that DMASP can reduce the accuracy of attacks against patterns leakage in real-world databases.