The Security Model of Unidirectional Proxy Re-Signature with Private Re-Signature Key

In proxy re-signature (PRS), a semi-trusted proxy, with some additional information (a.k.a., re-signature key), can transform Alice’s (delegatee) signature into Bob’s (delegator) signature on the same message, but cannot produce an arbitrary signature on behalf of either the delegatee or the delegator. In this paper, we investigate the security model of proxy re-signature, and find that the previous security model proposed by Ateniese and Honhenberger at ACM CCS 2005 (referred to as the AH model) is not complete since it does not cover all possible attacks. In particular, the attack on the unidirectional proxy re-signature with private re-signature key. To show this, we artificially design such a proxy re-signature scheme, which is proven secure in the AH model but suffers from a specific attack. Furthermore, we propose a new security model to solve the problem of the AH model. Interestingly, the previous two private re-signature key, unidirectional proxy re-signature schemes (one is proposed by Ateniese and Honhenberger at ACM CCS 2005, and the other is proposed by Libert and Vergnaud at ACM CCS 2008), which are proven secure in the AH model, can still be proven secure in our security model.