[arin-ppml] re-org question

Owen DeLong owen at delong.com
Sun Nov 6 16:08:07 EST 2016


> On Nov 6, 2016, at 12:47 , Matthew Petach <mpetach at netflight.com> wrote:
> 
> On Fri, Nov 4, 2016 at 7:55 PM, Owen DeLong <owen at delong.com> wrote:
>> You can return as small as a /24.
>> 
>> If you’re using half, then you can keep it.
>> 
>> So, at most, you have to renumber 126 hosts out of each of half of your /25s.
>> 
>> How is this not minimal again?
>> 
>> Owen
> 
> 
> I suspect Owen is trolling for effect here.

No, not really.

> 
> Renumbering 126 hosts out of 255 is only
> slightly less than half.

Sure, but that’s the absolute worst case and it only remains true if he
has managed to inefficiently use every single /24 in his entire address block.

> I'm guessing Owen would be unhappy, were he to be
> told by his doctor that he needed a surgical operation
> in which a minimal amount of his body mass would be
> removed; but upon delving deeper, found that the
> doctor would be removing slightly less than half his
> body.

It would probably depend on which pieces, but Owen is unusually obese and so
probably has more tissue than the average person he would be willing to part
with. Nonetheless, your point is made. However, see above… Using the absolute
worst possible case of the worst possible choices made in the past is probably
similarly atypical as an exemplar.

> Using the term "minimal" to apply to renumbering
> nearly half a block strays well past the realm of
> 'stretching the definition' into the neighborhood of
> 'trolling’.

But that’s also not likely to be the case.

> If the community does indeed think the language
> should stay, and that renumbering should be required,
> we should perhaps put some clarity around what is
> expected.  If an organization is using 40% of each /24,
> would the ARIN community be happy if the organization
> renumbered such that alternating /24s were now 80%
> filled, and returned every other /24 to ARIN, as individual
> /24 subnets?  That would meet the letter of the law, so
> to speak, but would ensure those blocks could never
> be aggregated into a larger allocation.  (As a side note,
> this could be a good way to ensure a steady supply of /24s
> for small entrants, while ensuring no larger entity can ever
> make use of them.)

I suspect that at this time, the community is unlikely to come to
a sufficient consensus in either direction and thus the existing
language remains despite multiple attempts to change it in either
direction.

> For the record, I think the sentence is confusing, and
> should have the "will" replaced with "may", to read as
> follows:
> 
> "ARIN will proceed with processing transfer requests even if the
> number resources of the combined organizations exceed what
> can be justified under current ARIN policy.  In that event, ARIN *may* work
> with the resource holder(s) to transfer the extra number resources to
> other organization(s) or accept a voluntary return of the extra number
> resources to ARIN.”

So you want it to be optional for ARIN to work with the resource holder,
but the resource holder is still required to return the space under section
12 and other parts of section 4? I am confused by this desire.

> emphasis on *may* included only to make it clear which word
> was changed; emphasis need not stay in the resulting NRPM
> text.  That should clear up confusion about it being a requirement,
> and instead make it clear this is an optional exercise.

Not the way you worded it, but at least now I understand your actual
intent. You’ll need to hack up a lot more of the NRPM than 8.2 to achieve
that desired result.

Owen





More information about the ARIN-PPML mailing list