Re: [DNSOP] Fwd: code points for brainpool curves for DNSSEC

Stephen,

Sorry for being so blunt below.

The document totally content free as to why this makes any sense in an
operational context.
DNSSEC algorithms should not be given out lightly as there is a significant
COST to deploy support for each additional algorithm.

While I strongly support having better ECC algorithm that the currently
specified curves, adding a new ones SHOULD take place via a performance
oriented process.
Thus I strongly advocate that this publication be held up until we can
compare this curve with other curves being proposed.

Background I'm currently fighting an multifaceted battle to have various
entities adding support for ECDSAP256, specified over 3 years ago.

Adding and deploying a new DNSKEY algorithm does not just require changes
to
-- DNS servers, libraries and resolvers.

That is just the top of the iceberg,  but also to
-  DNS provisioning systems, DNSSEC signing systems, DNS testing tools,
 - user interfaces for registrars, hosting providers, third party DNS
operators, CDN's,  etc.
 - TLD and ccTLD policy documents, EPP implementations, plus in some cases
security evaluations,
 - not to mention firewalls, network monitoring tools ....
 and number of other things I had no idea existed and majority of which is
not maintained anymore.

There are only so many times that one can get everyone's attention to
upgrade DNS stuff, thus requiring the change process to be run should not
be taken lightly.
If on the other hand if the editors are only interested in vanity algorithm
assignment without any deployment then ...........

Olafur

On Wed, Dec 9, 2015 at 4:00 PM, Stephen Farrell <stephen.farrell@cs.tcd.ie>
wrote:

>
>
>
> -------- Forwarded Message --------
> Subject: code points for brainpool curves for DNSSEC
> Date: Wed, 9 Dec 2015 18:00:18 +0000
> From: Stephen Farrell <stephen.farrell@cs.tcd.ie>
> To: saag@ietf.org
>
>
> Hiya,
>
> The brainpool folks have written an I-D [1] that they are pushing
> through the rfc editor's independent stream. [2]
>
> That I-D wants to register code points for using their curves for
> DNSSEC.
>
> For documents that come through the independent stream, the IESG
> does an RFC 5742 [3] conflict review. The purpose of that review
> is to check that the RFC doesn't conflict with ongoing work in
> the IETF.
>
> If you have thoughts on this, please let me know before Dec 17th.
> I'll forward this to the dnsop, cfrg and curdle mailing lists
> to check there too. Apologies if you get >1 copy of this. Please
> try follow up on the saag list if you can.
>
> Thanks,
> Stephen.
>
>
> [1] https://tools.ietf.org/html/draft-schmidt-brainpool-dnssec
> [2] https://www.rfc-editor.org/pubprocess/
> [3] https://tools.ietf.org/html/rfc5742
>
>
>
> _______________________________________________
> DNSOP mailing list
> DNSOP@ietf.org
> https://www.ietf.org/mailman/listinfo/dnsop
>