Verifiable top-k searchable encryption for cloud data

Abstract

With the proliferation of data and the appealing features of cloud computing, the data owners are motivated to outsource their data to the public cloud. Privacy and security, especially for sensitive data, are still a concern, as the data owners have no physical control over the outsourced data. To ensure confidentiality, sensitive data is encrypted before outsourcing to the cloud, which obsoletes the data utilization using traditional keyword-based search. To address this issue, a verifiable top-k searchable encryption for cloud data (VSED) is proposed with provisions for dynamic update operations like addition and deletion of documents. Specifically, an encrypted inverted index is constructed using a secret orthogonal vector and partial homomorphic encryption. To support the ranked search, the widely used term frequency and inverse document frequency rule is used to find the top-k documents. To verify the query results returned by the cloud server, this scheme provides a verifiable search using keyed hashes. Security analysis demonstrates that the proposed scheme is semantically secure, with correctness and privacy guarantees proved in the standard security simulation model. Simulations performed on real-world dataset demonstrate that the proposed scheme is efficient and practical.