Public-key Encryption with Keyword Search (PEKS) enables a server to test whether a tag from a sender and a trapdoor from a receiver contain the same keyword. In this paper, we highlight some potential security concern, i.e. a curious server is able to answer whether any selected keyword is corresponding to a given trapdoor or not (called an offline keyword guessing attack). The existing semantic security definition for PEKS does not capture this vulnerability. We propose a new concept, namely Public-key Encryption with Registered Keyword Search (PERKS), which requires a sender to register a keyword with a receiver before the sender can generate a tag. Clearly the keyword preregistration is not a pleasing feature. The payback is that the semantic security definition for PERKS proposed in this paper is immune to the offline keyword guessing attack. We also propose a construction of PERKS and prove its security. The construction supports testing multiple tags in batch mode, which can significantly reduce the computational complexity in some situations.
|Title of host publication||Sixth European Workshop on Public Key Services, Applications and Infrastructures|
|Place of Publication||London|
|Number of pages||21|
|Publication status||Published - Sep 2009|
|Name||Lecture Notes in Computer Science|
- offline password guessing attack
Tang, Q., & Chen, L. (2009). Public-key Encryption with Registered Keyword Search. In Sixth European Workshop on Public Key Services, Applications and Infrastructures (pp. 163-178). (Lecture Notes in Computer Science; Vol. 6391). London: Springer. https://doi.org/10.1007/978-3-642-16441-5_11