Re: [Curdle] I-D Action: draft-ietf-curdle-ssh-ed25519-01.txt

[denis bider <[e-mail]>]

Aye, you're right. That is a concern for us as well. SUPERCOP has a
MIT-licensed implementation from Mike Hamburg, but this would be a problem
for us to, because it requires updating our license terms just for this
algorithm.

I have historically avoided any external code that has an impact on our
license terms. In the absence of a public domain implementation, I would
have to either change this policy, or implement my own. And I don't want to
subject anyone to crypto I implement on my own. :)

On Sun, Aug 13, 2017 at 3:38 AM, Damien Miller <[e-mail]> wrote:

> On Fri, 11 Aug 2017, Mark D. Baushke wrote:
>
> > Hi denis,
> >
> > Yes, the lack of two implementations is a potential problem.
> >
> > Given the libssh.org folks did the original work for curve25519,
> > I wonder if they will be the first to do curve448 and likewise eddsa448.
>
> I looked at adding ed448 to OpenSSH a while back and got stuck looking for
> a standalone ed448 implementation that was as small, self-contained, clean
> and suitably licensed as the ed25519 that we use (from Supercop).
>
> Now, IMO our priority should be to choose a PQ KEX before another EC one.
>
> -d
>