[arin-ppml] IPv4 Transfer Policy Change to Keep Whois Accurate

[Owen DeLong]

On May 12, 2011, at 11:48 AM, Matthew Kaufman wrote:

> 
> On May 12, 2011, at 10:02 AM, Owen DeLong wrote:
> 
>> 
>> On May 12, 2011, at 8:23 AM, Matthew Kaufman wrote:
>> 
>>> On 5/12/2011 2:57 AM, Owen DeLong wrote:
>>>> On May 11, 2011, at 7:41 PM, Mike Burns wrote:
>>>> 
>>>> 
>>>>> There ARE other addresses available, I have heard that you are aware of IPv6.
>>>>> 
>>>> Sigh... Even I am not going to attempt to claim that IPv4 and IPv6 addresses are
>>>> interchangeable equivalents.
>>>> 
>>> 
>>> I keep hearing arguments that the price of IPv4 addresses won't go infinitely high because at some price it'll be more cost-effective to switch to IPv6 than to try to keep shuffling the IPv4 deck chairs.
>>> 
>> 
>> You haven't heard that argument from me and I'm not convinced it is accurate.
> 
> Ok, so what do *you* think is going to happen if the price of IPv4 addresses converges with infinity?
> 
MOAR NAT.

Unfortunately.

Probably it will increase the pressure towards IPv6 marginally, but, not to such an extent on both sides
of the equation as to accomplish much in and of itself. Worse, if enough incumbent Telcos/Cablecos
get enough addresses, it might even serve to forestall their considering IPv6 a necessary step in favor
of using that fact to bludgeon their smaller competitors into submission.

>> 
>>> If that is true, then IPv4 and IPv6 addresses *are* economically interchangeable, just as coal is an imperfect substitute for natural gas when it comes to generating electric power... at some point the cost of one or the other gets high enough that the cost of switching is lower.
>>> 
>> 
>> Since it may not be true... 
>> 
>> With coal vs. CNG, you have a single party able to make the switch deterministically for themselves.
>> With IPv4 vs. IPv6, you have a myriad of third-party dependencies in your ability to switch from
>> one to the other and have it actually work out. This significantly changes the underlying assumptions.
>> 
>>>>> And with my policy or without my policy, addresses are bound to flow to the highest bidder.
>>>> If the highest bidder is limited to only those addresses he can justify, then the addresses he
>>>> couldn't justify flow to the next highest bidder. If the highest bidder is not so limited, then, the
>>>> addresses likely flow to a very small number of very well capitalized entities to the extreme
>>>> detriment of smaller entities.
>>> 
>>> The addresses will likely flow to a very small number of very well capitalized entities in any event... the only question is what type of entity they are. In Mike's world they flow to folks who do things like lease address space and sell blocks at high prices to people who really need them... in your world they flow to the top N ISPs that are experts at showing need, aren't constrained by the 3-month rules, and are growing sufficiently to justify anything.
>>> 
>> 
>> I think the results are actually the top N ISPs in both cases eventually. The difference is that Mike's resultant value
>> of N tends to be much smaller than mine.
> 
> I'd argue the reverse, actually. If there is a needs basis then the top N ISPs have an advantage in that they are already experts at manipulating the needs basis. (And it really is manipulation, because any of them could actually get by with much much less than they already have if the "need" took into account things like forcing customers to use NAT)

Forcing customer to use NAT is not within the scope of ARIN policy and I do not consider the failure or refusal to do so to be a manipulation of need.
NAT is a degraded subset of internet access services.

There are many providers of much smaller size that are equally expert at dealing with ARIN justified need. I do not believe your argument
hold water. As I said, I think with needs-basis, the value of N tends to be larger than it will be without N.

> 
> One could argue, for instance, that *with* a needs basis Comcast might end up holding nearly all the space... but without a needs basis it might be an investment banking firm instead, who'd then lease the space out to move providers than just Comcast.
> 

I think you have a number of assertions there which are not at all proven:

	1.	I think it is unlikely $CABLECO could demonstrate 12 month need for all the
		addresses on their first go. Other providers and end-users, will therefore be
		likely able to obtain some fraction of the address space.

	2.	I have no reason to believe that $CABLECO would not be ahead of the
		investment banking firm in line for the addresses as I suspect $CABLECO
		generally pays more attention to the goings on at ARIN than does
		$INVESTMENT_BANK.

	3.	There is nothing to prove that $INVESTMENT_BANK would not prefer to
		lease all the addresses to $CABLECO vs. dealing with multiple tenants.
		In fact, there are many things to suggest that such a model would be
		highly preferable as it would maintain roughly the same revenue while
		simultaneously reducing costs which generally is viewed favorably
		by most of the investment bankers I know.

Indeed, I think that the likelihood is each of those three assertions is not correct.

>> 
>>> In Mike's world, you can get service from whoever you want but the price of space is high... in your world the price of (what is now provider-assigned) space is high and your choice of transit providers is limited.
>>> 
>> 
>> I really don't see how you come to these conclusions. My argument is that under Mike's proposed policy,
>> your choice of providers would become more limited.
> 
> I think you've got it backwards... see above.
> 

No, I am not the one who has it backwards. See above.

Owen