One-time encryption-key technique for the traditional DL-based encryption scheme with anonymity

In modern cryptosystem, Anonymity means that in some sense any adversary cannot tell which one of public keys has been used for encrypting a plaintext, and was first formally defined as the indistinguishability of keys by Bellare et al. in 2001. Recently, several well-known techniques have been proposed in order to achieve the anonymity of public-key encryption schemes. In this paper, anonymity is considered first from a new perspective. And then basing on this new perspective, a one-time encryption-key technique is proposed to achieve the anonymity of traditional discrete-logarithm-based (DL-based) encryption scheme. In this new technique, for each encryption, a random one-time encryption-key will be generated to encrypt the plaintext, instead of the original public-key. Consequently, in roughly speaking, by the randomness of the generated one-time encryption-key, this new technique should achieve the anonymity. Furthermore, in the formal proof of anonymity, only based on several weaker conditions, the one-time encryption-key technique efficiently achieves the provable indistinguishability of keys under chosen ciphertext attack (IK-CCA anonymity). As a result, compared with the work of Hayashi and Tanaka in 2006, the one-time encryption-key technique presented here has fewer requirements for achieving the provable anonymity.