Reading some questions in this forum I encountered this answer mentioning IND-CPA as a requirement for ciphers. It got me thinking about this paper I recently read, where a formal model of the Honest-But-Curious adversary is given, in order to automate the analysis of whether certain protocol has the indistinguishability property. And I have this way realized I am missing a piece in the puzzle.
As far as I remember, in a Cryptographic System, we have:
- Ciphertext-only (Passive Attack)
- Known Plaintext Attack (Passive Attack)
- Chosen Plaintext Attack (CPA)
- Chosen Ciphertext Attack (CCA)
And Notions of Security like
- Semantic Security (SS, SEM)
- Indistinguishability Security (IND)
- Non-malleability (NM)
And for stating the notions preserved under certain attacks we end up with combinations like the well known IND-CPA.
Now, how do adversarial models like Dolev-Yao, Honest-But-Curious or Byzantine fit in this scheme? ( If they do at all and I am not mixing up things)