Re: [dnsext] afasterinternet.com trial and draft-vandergaast-edns-client-subnet-00

[Ted Hardie <[e-mail]>]

On Wed, Aug 31, 2011 at 10:30 AM, Andrew Sullivan
<[e-mail]> wrote:
> On Wed, Aug 31, 2011 at 09:03:34AM -0700, Ted Hardie wrote:
>> again, or is it your opinion that the code point allocation in an Experimental
>> RFC does not need to address these issues?
>
> Olafur and I consulted over this, and here is our opinion as chairs.
> Note that we've been wrong before.
>
> As a pure matter of the rules, it is our opinion that the code point
> allocation does not need to address any objections to the idea.  The
> rule is only "RFC Required".  An Experimental track RFC is an RFC, and
> therefore it would qualify.  One might take issue with the experiment
> itself, and one might take issue with what the experiment is
> attempting to prove, but none of that would be reason to refuse to
> publish the RFC nor to deny the allocation of a code point.
>
> So, any RFC will do.  Even an RFC that simply says, "IANA
> considerations: IANA allocate EDNS0 option NETMASK code TBD.  The
> option is a text string of one of the following formats:
> 'IP4_addr/mask' or 'IPv6_addr/mask'", along with perhaps a section
> that says by way of introduction, "This document specifies the NETMASK
> EDNS0 option, but does not specify its use."
>
> Taking my hat off, I note that, if this opinion is correct, RFC
> Required appears to become a lower bar than Expert Review.  I am not
> sure that the IESG would agree with our interpretation, therefore.
>
> A

Traditionally expert review has been a lower bar than RFC required
because you have to convince someone that it's a good idea to publish
the RFC (The independent stream editor, some working group, some AD).
That's not been trivial, and it usual comes with questions about the
track.  An experimental RFC, for example, might be tasked with
describing the experiment and what results would constitute success.
I can't really see the internet-draft text you cite above passing IETF
last call without a few questions.  We have seen some docs where the
registration document is recognition of a de facto use that is being
codified to avoid collision, but even those describe what's going on.
Doing that too often with code points that aren't unlimited has, of
course, obvious failure conditions.

You and Olafur previously said (as individuals)  "We stand ready and
willing to help
you pursue that publication, particularly since you intend your
protocol to be experimental.".
Since this was evidently not through the working group, did you intend
AD sponsorship or an
independent stream submission?

regards,

Ted Hardie