ACP-lrFEM: Functional Encryption Mechanism with Automatic Control Policy in the Presence of Key Leakage

摘要

We present a leakage-resilient functional encryption from finite automata control policy, in which the ciphertext is associated with an input string w and the private key is connected to a finite automata M. The decryption will succeed iff the automata accepts the string, i.e., Accept(M,w) = 1. In our scheme, we allow the leakage of sensitive key by allowing the attacker to provide an efficiently computable function (leakage function) adaptively, and to receive the output of the function taking the private key as input. Our security model considers two sides: key-leakage resilience and plaintext confidentiality. Not only can the attacker request the reveal of all non-match keys of finite automata, but can query the leakage for the match key. We also deploy an update algorithm to support the continual leakage resilience. We give the construction in bilinear groups of composite order and prove the security in dual system framework. The analysis shows that the maximum leakage of the key can be 33%.