In reference to the LR-Oracle experiment in “Introduction to Modern Cryptography” (2nd edition) by Lindell & Katz, Definition 3.23 states a scheme $\pi = (Gen,Enc_K,Dec_K)$ is CPA secure for multiple encryptions if (to use the notation of Lindell and Katz)

$$\begin{equation} \textbf{Pr}[\textbf{PrivK}_{\mathcal{A},\pi}^{LR-cpa}(n) = 1] \leq 1/2 + \textbf{negl(n)}\end{equation}$$

Some questions:

  1. I'm assuming that this implies that for $\pi$ that satisfies the above, it's safe to provide oracle access to $Enc_{K}(\cdot)$ to a CPA adversary $\mathcal{A}$, and by implication its safe to expose a counter mode encryption oracle to a CPA adversary

  2. If not, what is the exact definition that tells me the conditions under which it's safe to expose an encryption oracle to a CPA adversary?

  3. Since CCA security requires non-malleability in the cipher, is it true to say that adding authentication to a CPA secure scheme in the manner of an AEAD scheme is the sole requirement to additionally provide non-malleability / security against a CCA adversary, thereby making it safe to additionally expose a decryption oracle to $\pi$ to the adversary?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.