Re: Genart last call review of draft-ietf-curdle-ssh-ed25519-ed448-07

[Viktor Dukhovni <[e-mail]>]

On Sun, Jan 06, 2019 at 07:50:40PM -0800, Linda Dunbar wrote:

> This document proposes two names for public key algorithms (which are specified
> by other RFCs): ssh-ed25519 & ssh-ed448
> 
> Major issues:
> 
> There is no "Standard" being specified by this document. The document has a few
> sentences to explain "public key algorithm for use with SSH in accordance with
> RFC4253, RFC4251" and give a name. and One sentence to say "Signatures are
> generated according to the procedure in RFC8032".

I've been doing some work on a fork of OpenSSH 8.0p1 lately, enhancing
GSS Key Exchage support, adding modern GSS KEX algorithms, ... and from
my perspective there is indeed a relevant standards action here, namely
this defines a "standard meaning" of those names in the context of SSH
to be the use of the corresponding hostkey algorithms (where the
mechanics of computing the associated signature are already defined
elsewhere).

FWIW, IMHO the new gss key exchange algorithms (gss-group14-sha256-*,
gss-curve25519-sha256-*, ...) in draft-ietf-curdle-gss-keyex-sha2-10
are similary worthy of that draft's "Standards Track" designation.
I have assumed that draft will also become an RFC in the not too
distant future, and those code points are already implemented in
my fork.

-- 
	Viktor.