[ppml] Policy Proposal: Resource Reclamation Incentives

[michael.dillon at bt.com]

>   - Specifically, that their space cannot be revoked for 
> failing to use
>     it.

That is going to be intractable as we get closer to IPv4 exhaustion.
When ISP X is facing losses of millions of dollars due to the fact that
ARIN has no more addresses to give, and it is known that ISP Y has an
unused hoard, then what happens.

Perhaps ISP X simply "borrows" ISP Y's hoard. Will Y go to court and
argue that X can't do that because ARIN's whois shows that these are Y's
addresses? Will X argue that Y has no right to addresses because they
are not an ARIN member and have no contract with ARIN, therefore no
expectation of service from ARIN? Will the court revoke all of Y's
addresses? Will the court order Y to return unused addresses to ARIN
because they are, by proxy, causing ARIN to materially damage X's
business?

The complexity makes your head spin. In general, courts are supposed to
be a last resort after the parties have sat down, discussed the issue,
and tried to hammer out an agreement. Can Y really be said to have done
this if Y has not joined ARIN, signed the RSA, returned excessive unused
addresses and made and honest attempt to comply with ARIN policy? ARIN
is not just a 3rd party here; ARIN is the forum in which all parties
come together and hammer out the mutual agreement to shared these
limited resources.

Organizations which do not enagage with each other in ARIN are the
rogues who refuse to negotiate.

It's hard to predict the legal outcomes but one thing is certain, any
organizations which persist in refusing to join ARIN will suffer the
consequences of a lot of negative publicity as we get closer to IPv4
exhaustion. Even if they don't don the black hat and get involved in a
court case, these organizations will be tarred with a nasty brush. In
particular the larger ones with class A and B allocations because one
would expect them to have the most capability for implementing IPv6 and
reducing their consumption of IPv4 addresses. 

> - Turn in your existing bucket of disjoint small netblocks for
>   one large, aggregateable netblock.

Hmmm... Thought experiment.
Entity A has acquired 300 class C address blocks. None of these are used
at all, just acquired. Entity A goes to ARIN and turns in their 300
class C's in return for a /16 which is 256 /24 equivalents, therefore it
is less than their original hoard. Now Entity A markets itself to
organizations who are concerned about the pending IPv4 shortage. They
manage to rake in 6 figures for their shell company and the buyer gets a
/16 to use. Sale is timed to close shortly after the buyer gets their
next ARIN allocation.

Obviously this could only occur if Entity A does not have to fully
justify their space. Currently the wording of the policy does allow for
this loophole.

I can't help but think that we would make better policy if we started
out with use-cases and requirements rather than diving straight into the
specific text of a policy.

--Michael Dillon