Insight of the protection for data security under selective opening attacks

Data security and privacy protection issues are the primary restraints for adoption of cloud computing. Selective opening security (SOA security) focuses on such a scenario of cloud computing: Multiple senders encrypt their own data with the public key of a single receiver. Given the ciphertexts, the adversary is allowed to corrupt some of the senders, seeing not only their plaintexts but also the random coins used during the encryption. The security requirement of SOA security is that the privacy of the unopened data is preserved. On the other hand, non-malleability is also a very important security notion for data security in cloud computing and public-key cryptography. The security requirement of non-malleability is that given a challenge ciphertext, it should be infeasible to generate a ciphertext vector whose decryption is "meaningfully related" to the corresponding challenge plaintext. However, as far as we know, the relations between non-malleability and SOA security are still undiscovered, and the security notion of non-malleability under selective opening attacks has not yet been formally defined or researched. In this paper, we formalize the security notion of non-malleability under selective opening attacks (NM-SO security), and explore the relations between NM-SO security and the standard SOA security, the relations between NM-SO security and the standard non malleability, and the relations among NM-SO security notions. (C) 2017 Elsevier Inc. All rights reserved.