[arin-ppml] re-org question

Owen DeLong owen at delong.com
Fri Nov 4 22:45:50 EDT 2016

> On Nov 4, 2016, at 19:33 , William Herrin <bill at herrin.us> wrote:
> On Fri, Nov 4, 2016 at 10:28 PM, Owen DeLong <owen at delong.com> wrote:
>>> On Nov 4, 2016, at 19:11 , William Herrin <bill at herrin.us> wrote:
>>> On Fri, Nov 4, 2016 at 10:01 PM, Owen DeLong <owen at delong.com> wrote:
>>>> Perhaps, in part, because some of us think that the RSA is what is broken rather than the language in the policy.
>>> Owen,
>>> Really? You think ARIN policy should be that folks are required to
>>> renumber just because of a business transaction involving the owner of
>>> the physical network using them?
>> Nowhere does it say you are required to renumber. You’re reading that into things.
>> It says that if the combined organization can not justify the combined resources, that ARIN will work with them on appropriate transfers, and/or returns.
> Owen,
> You're going to have to explain that particular magic to me because I
> haven't heard of any way to free up netblocks for return other than
> renumbering out of them.

If you’re using 50%, you pretty much meet current guidelines, so really very little problem.

If you’ve got a /16 and you’ve got /28s all over it, yeah, you’ll probably have to do some consolidation (renumbering) of some of the networks.

If you’ve got a /16 and you’re using most of the first /18, then no problem with returning the last /17 and (if needed) the second /18.

> Sure, an org might possibly have vast amounts of unrouted address
> space they're just not using with all the addresses they are using
> carefully crammed into an efficient /24 but what's your plan for
> everybody else?

There’s a lot of ground in between those two extremes. In the vast majority of cases I’ve encountered, minimal renumbering can free up more than
enough space to return to satisfy policy.


More information about the ARIN-PPML mailing list