Re: [Gen-art] review of draft-os-ietf-sshfp-ecdsa-sha2-04.txt

[Francis Dupont <[e-mail]>]

 In your previous mail you wrote:

>  > Minor issues: not a real issue but I am not convinced there is a real
>  > crypto reason to give up SHA-1. At the first view the attack against
>  > SSHFP is a pre-image one, but:
>  > - I leave the question to cryptographers of the security directorate
>  > - there are many not-crypto reasons to move from SHA-1 to SHA-256
>  
>  Hi,
>  
>  I have added some text there:
>  
>            ECDSA public key fingerprints MUST use the SHA-256 algorithm
>            for the fingerprint as using the SHA-1 algorithm would
>            weaken the security of the key, which itself can use only
>            SHA-2 family of algorithms RFC 5656 (Section 3.1.1).

=> I am afraid it is another not-crypto reason...

>  But I am also not a cryptographer,

=> I am not a cryptographer too (I just worked with cryptographers,
military cryptographers exactly, i.e., the worst kind of
cryptographers :-)

>  so it's just my guts telling me
>  that if a key is allowed to use only SHA-2, we should keep it in sync
>  here.

=> the 2 ideas are:
 - keep the requirement (i.e., it is the right one and even there could
  be no good crypto reasons)
 - get a wording for the justification which doesn't make cryptographers
  too unhappy (they won't be really happy anyway: this is a part of
  being cryptographers :-)

Of course for the second part the best should be to get a feedback
from the crypto (oops, the security) directorate.

Thanks

[e-mail]