ABE with improved auxiliary input for big data security

Attribute-based encryption (ABE) is recommended by the Cloud Security Alliance (CSA) as one of the possible cryptographic tools for access control in big data applications. In ABE, the shared file can be encrypted with the specific policy only once, and it can be decrypted by any receiver whose attributes are satisfied. When ABE is deployed in some open network scenarios, it is inevitably attacked by side channel attacks, because the big data are coming from diverse end-points. In this paper, we propose leakage resilient CP-ABE and KP-ABE schemes in the improved auxiliary input model, which allows the attacker query more leakage information regarding the encryption randomness after seeing the challenge ciphertext. Moreover, we construct an improved strong extractor from the modified Goldreich-Levin theorem for the security proof and prove that our scheme security relies on the Wang et al. construction.