Privacy-preserving and adaptively-secure encryptions with Deterministic Finite Automata policy and their applications

Deterministic Finite Automata (DFA) is a useful mathematical tool in defining the finite states and describing the transition among these states. In this paper, we propose two adaptively secure functional encryptions in the standard model that are based on DFA policies. In the first scheme, the ciphertext is associated with a DFA M and the token is associated with an arbitrary length string w, and there is a check algorithm to test whether the string w is accepted by the automata M in the key/ciphertext spaces. In the second scheme, we extend the first scheme to support payload confidentiality, in which the decryption can extract the encrypted message if the associated automata accepts the string. Using the technique of dual system encryption, we prove the schemes can achieve adaptive security under the static assumptions, and we then give the performance evaluation. We also provide the deployments in privacy-preserving outsource computation in cloud, text filtering in firewall, and privacy-carrying DNA match in decentralized network etc.