Re: ECB-Counter AES mode

Karl Malbrain <malbrain@xxxxxxxxx> wrote:

Yes, taking a sequential counter and encrypting it (e.g. CTR mode) then
encrypting the CTR mode result would avoid the problems with plain-old CTR
mode.

Huh? What problems with plain-old CTR mode? CTR mode is IND-CPA with a
secure PRP. The only problems I know of are chosen-ciphertext attacks
and lack of plaintext integrity. These problems have the same solution:
authenticate the ciphertext.

Re-encrypting the CTR-mode output provides a minimal level of protection
in the case where you're sufficiently stupid to reuse a counter value.

-- [mdw]
.

Quantcast