Re: [Idr] Working group adoption call for draft-snijders-idr-deprecate-30-31-129

[Peter Hessler <[e-mail]>]

On 2016 Nov 01 (Tue) at 15:17:30 +0100 (+0100), Robert Raszuk wrote:
:> An implementor should not be rewarded with the possibility
:> of laziness after they squat on a codepoint.
:
:+
:
:> I argue that we cannot tolerate:
:
:Ahh got it.
:
:So the point of the document is *punishment* for early attempts to
:implement anything not to technically address the problem at hand. Thx for
:clarification ...
:
:Cheers,
:R.
:

Early development in support of an ID is desirable, and in this WG
pretty much required.  For that, a lab image and (in the open source
case) patches floating around is good.  Pushing these to innocent
customers is not.

Squatting on random (e.g. Huawei with code 30) codepoints *should* be
punished, imho.

Regardless if we consider it to be punishment or not-punishment, it
should be documented that these code points are tainted, and cannot be
considered pure and clean for future assignments.

I don't care either way if 129 is used for a future Early Assignment of
Wide Communities.  But I do care if 31 is assigned to someone that has
to fight the existing ecosystem.



:On Tue, Nov 1, 2016 at 3:12 PM, Job Snijders <[e-mail]> wrote:
:
:> On Tue, Nov 01, 2016 at 03:02:37PM +0100, Robert Raszuk wrote:
:> > Reason being that if any of the applications they have been squatted
:> > on behalf of want to use them - it would be pretty easy to allocate
:> > them.
:> >
:> > The other alternative would be to allocate them to those applications
:> > now (for which the drafts exist) as IANA early allocations. If you
:> > depreciate them now and we know that as example Contrail Multicast is
:> > important either your draft will need to quickly be moved to historic
:> > or yet another types will need to be allocated again.
:>
:> That is not how the process works. An implementor should not be rewarded
:> with the possibility of laziness after they squat on a codepoint.
:> "Laziness" being that they don't have to renumber, and that they didnt
:> have to go through the normal process to obtain the codepoint.
:>
:> I argue that we cannot tolerate: "I'll just take this one, IANA will
:> give it to me anyway later on if the product becomes popular"
:>
:> Also, the 'deprecate' draft (if published) would be "updated", not moved
:> to historic. Only when all six proposed squatted points are returned to
:> the "unassigned" pool the document can be declared historic.
:>
:> > In any of the above Large Communities will get virgin BGP attribute
:> > type which I believe is the main point here.
:>
:> No, the main point is to prevent another deployment issue like Large BGP
:> Communities suffered. Large BGP Communities has been resolved now (after
:> considerable effort I might add).
:>
:> This document protects innocent bystanders from being assigned
:> undeployable tainted codepoints.
:>
:> Kind regards,
:>
:> Job
:>


-- 
Command, n.:
	Statement presented by a human and accepted by a computer in
such a manner as to make the human feel as if he is in control.