IND-RCCA Security

The basic definition of Indistinguishability under Replayable CCA.

Replayable CCA

Definition

An encryption scheme $\Pi = (\Gen, \Enc, \Dec)$ is IND-RCCA-1 secure if for all PPT adversaries $A:=(A_1, A_2)$, $\Game_{\Pi, A}^{\textrm{RCCA}}( \lambda, 0) \approx_c \Game^{\textrm{RCCA}}_{\Pi, A}(\lambda, 1)$, where the game is the following:

\[\begin{aligned} &&& \underline{\Game_{\Pi, A}^{\textrm{RCCA}}( \lambda, b)}: \\ & \\ &1. && (pk, sk) \samples \Gen(1^\lambda) \\ &2. && (m_0, m_1, z) \samples A_1^{\Dec(sk, \cdot)}(pk) \\ &3. && c \samples \Enc(pk, m_b) \\ &4. && b' \samples A_2^{\Dec^*(sk, \cdot)}(c, z) \end{aligned}\]

The adversary has oracle access to $\Dec^*$, which returns a decryption $m$ of the input chipertext $c$; if $m \in {m_0, m_1}$, then the oracle returns a special symbol. This is the main difference with respect to the CCA scenario.