[arin-ppml] ARIN-2011-11 and ARIN-2011-12 - Last Call

[John Curran]

On Jan 12, 2012, at 5:11 PM, Michael Sinatra wrote:
> 
> The 3-month need only applies to the ARIN free pool, not to 8.3 transfers, as NRPM Section 4.2.4.4 notes.  8.3 transfers already get a 12-month need window.
> 
> The difference in the two need windows further distorts the market for IP addresses, unnecessarily inflating transfer market prices.  It also prolongs the pathological situation where ARIN has a large free pool relative to other regions, presents the impression (wrongly, I agree) that ARIN is hoarding IP resources, and it places unnecessary burdens on ISPs attempting to legitimately get IP address space.  Any concern that the ARIN pool would be raided by out-of-region entities is both a product of the current situation that shouldn't be prolonged and something which can be mitigated by placing time limits on transfers of recently-acquired address space.
> 
> The free pool should match the transfer pool, and the free pool need should be liberalized, rather than the transfer pool being made more strict.  As long as there are free addresses, ARIN should not attempt to protect them by giving incentives to use the transfer market and disincentives to come to ARIN for IPv4 resources.
> 
> Until there is an effort to normalize the free pool and transfer policies, I can't support 2011-12, which would only increase the current distortions.

Micheal - 
 
Under the present policies, we are indeed seeing established service
providers looking to obtain addresses from the transfer market due to
the certainty of getting up to 1 years worth of their documented need.  
Since the current allocation policy only provides for obtaining space
equal to 3 months of demand directly from ARIN's free pool at a time, 
it is not clear to some providers that they'll be able to rely on the 
free pool in the region to obtain the equivalent address space.

If address space received by direct allocation from the regional 
free pool equalled what could be obtained via transfer from existing 
address holders, then indeed there'd be very little reason for a 
party in the region to obtain addresses via transfer (which would 
be a desirable goal from your viewpoint if I understand your messages 
correctly.)  It also would provide for similar routing impact when 
allocated address blocks are introduced into the global routing table,
as opposed to the present situation involving more smaller blocks 
allocated for the same total space received.

When one looks at the history of the 3 month allocation policy, it 
was considered a way of preventing a disproportionate advantage for
service providers who would otherwise receive 12 months of space 
near the end of the free pool when compared those who immediately 
follow.  Do you feel that the benefits of limiting that advantage 
at runout are not worth the distortions creating with respect to 
the specified transfer policy?  I ask solely to better understand 
your perspective, since there are interactions between the various 
policy goals here and the relative tradeoffs between them have to 
be considered.

Thanks!
/John

John Curran
President and CEO
ARIN