Verifiable Searchable Symmetric Encryption Over Additive Homomorphism

Abstract

Searchable symmetric encryption (SSE) allows the client to search encrypted documents on an untrusted server without revealing the document content and queried keywords. To improve search efficiency and enrich expressiveness, most SSE schemes leak some information that could be exploited for attacks, characterized by leakage patterns. The traditional leakage patterns encompass the search pattern, the access pattern and the response length pattern. Recent research has demonstrated that these three patterns could be exploited to launch attacks, resulting in a high probability of compromising the confidentiality of encrypted documents and queried keywords. Moreover, while there exist SSE schemes that hide multiple leakage patterns, most of them do not resist the malicious server, which may carry out incorrect search operations. In this paper, we propose a leakage-suppressed verifiable SSE (VSSE) scheme that not only hides the three patterns but also allows the client to verify the server’s response. We utilize the privacy set intersection based on polynomial coding and additive symmetric homomorphism encryption to construct a VSSE scheme that supports a conjunctive query. Specifically, we design an efficient random token generation algorithm to protect the search pattern and a verification algorithm that does not require server-generated proofs. Formal security analysis shows that our scheme achieves the desired correctness, security and verifiability. Lastly, we simulate the proposed scheme and compare it with the recent leakage suppression schemes in multiple aspects. The comparison results show that our scheme achieves a good balance in expressiveness, efficiency and security.