Functional Encryption against Probabilistic Queries: Definition, Construction and Applications
Functional encryption (FE for short) can be used to calculate a function output of a message, without revealing other information about the message. There are mainly two types of security definitions for FE, exactly simulation-based security (SIM-security) and indistinguishability-based security (IND-security). The two types of security definitions both suffer from their own drawbacks: FE with SIM-security supporting all circuits cannot be constructed for unbounded number of ciphertext and/or key queries, while IND-security is sometimes not enough: there are examples where an FE scheme is IND-secure but not intuitively secure. In this paper, we present a new security definition which can avoid the drawbacks of both SIM-security and IND-security, called indistinguishability-based security against probabilistic queries (pIND-security for short), and we give an FE construction for all circuits which is secure for unbounded key/ciphertext queries under this new security definition. We prove that this new security definition is strictly between SIM-security and IND-security, and provide new applications for FE which were not known to be constructed from IND-secure or SIM-secure FE.