[arin-ppml] Draft Policy 2012-4: Return to 12 Month Supply and Reset Trigger to /8 in Free Pool

[Owen DeLong]

On Mar 14, 2012, at 7:54 PM, Joe Maimon wrote:

> 
> 
> Owen DeLong wrote:
>> I want ARIN to be able to continue to hand out resources indefinitely.
>> 
>> I don't see that changing based on this policy.
>> 
>> The policy scope applies only to IPv4 resources. For those, it is not a question of how long we want resources to be available, but, how we distribute the pain of runout.
>> 
>> Owen
> 
> How long we want ipv4 resources available from ARIN is very much the exact question.
> 
> I know you read the staff assessment, where that is stated explicitly.

My point is that convincing ourselves that we can extend that date is an act of fiction. In reality, we're already past runout and we have been for years if you define runout as the point at which genuine needs cannot be fully and rationally satisfied. We passed runout as far as I'm concerned at the point where we were no longer able to take every host that is currently behind NAT and put a public address on it instead.

We've been out of IPv4 resources for years, but that's an inconvenient truth. So, instead we find new and different ways to lie to ourselves and pretend that the free pool today is not measured in the pain of others and that we can magically extend the life of that free pool, conveniently ignoring the pain such actions inflict.

There is already pain from IPv4 runout reflected in current policy. Anything we do to extend the imaginary life of the free pool comes at the price of increasing that pain for all who are currently bearing it and for others who will bear it in the future as well.

Hence, I stand by my statement that it is not about extending the life of the free pool, but, rather about how we distribute the pain of runout and how long we keep that pain going.

If the window stays at 3 months, it's not like those large requests don't still represent real need. The large requests are either spread out over 4 requests or we have managed to make the pain/cost of address acquisition exceed the value for those requestors (possibly to the expense and detriment of their customers). Essentially, we have robbed $LARGE_REQUESTOR to provide resources to n * $SMALL_REQUESTORS where the value of n is unknown and n may actually be entirely made up of smaller requests from the same organization resulting in greater fragmentation of the routing table while not changing actual address consumption.

Owen

> 
>>>>  Staff Assessment.
>>>> 
>>>> • Issuing a 12-month supply of IPv4 addresses will likely significantly
>>>> accelerate the depletion of ARIN’s existing IPv4 free pool.
>>>> Historically, ARIN’s IPv4 consumption rate was roughly doubled when
>>>> issuing a 12-month supply vs a 3-month supply.
>>>> o From 2008 through 2010, ARIN issued 3.36, 2.46, and 2.69 /8s
>>>> respectively when issuing a 12-month supply, vs 1.32 /8s in 2011 when
>>>> the 3-month supply policy went into effect.
>>>> 
>>>> • With the reintroduction of a 12-month supply window, there is the
>>>> possibility that several very large requests could quickly deplete
>>>> ARIN’s free pool. In light of this fact, the community may want to
>>>> consider bringing back a maximum allocation/assignment size.
>>> How long do you want ARIN resources to be generally available for?
>>> 
>>> 1-2 years?
>>> 
>>> 3-5 years?
>>> 
>>>