[ppml] Policy Proposal 2007-14: Resource Review Process

[Stephen Sprunk]

Thus spake "Bill Woodcock" <woody at pch.net>
>    > 4. If the review shows that existing usage is substantially not in
>    > compliance with current allocation and/or assignment policies,
>    > the organization shall return resources as needed to bring them
>    > substantially into compliance. If possible, only whole resources
>    > shall be returned. Partial address blocks shall be returned in
>    > such a way that the portion retained will comprise a single
>    > aggregate block.
>
> How do you envision this working, in the case that someone is
> both lightly and sparsely using the space they've been allocated?
> That is, if someone is using the first and last addresses of a /22,
> would they be forced to renumber, in order to create a returnable
> /23 from one end or the other?

Presuming returning a /23 would be sufficient given how they're using the 
/22, yes.

Consider a pathological case where someone has a /16 but is only using every 
fourth /24.  If they were allowed to return the parts they weren't using, 
that'd result in a return of 64 /23s and 64 /24s, and they'd be forced to 
advertise the remaining 64 /24s.  That's bad for everyone -- including the 
returning org.  Odds are they wouldn't be able to get their /24s past 
filters anyways, and the other /23s and /24s would be equally unusable if 
ARIN tried to reassign them to someone else.

(Actually, I'd expect in that case they'd return 64 /23s and get to keep the 
other unused 64 /24s, meaning they'd be advertising 64 /23s.  That's 
irrelevant, though, since they'd hit the same filters.  Much better that 
they renumber into one /17 and return the other.)

S

Stephen Sprunk      "Those people who think they know everything
CCIE #3723         are a great annoyance to those of us who do."
K5SSS                                             --Isaac Asimov