[arin-ppml] Draft proposal that needs some wordsmithing

[Mike Burns]

Hi Ray,

>> After exhaust a free market is the appropriate mechanism to ensure 
>> addresses
>> are allocated and used efficiently.
>>
> A claim completely unsupported by any credible evidence whatsoever.
>
> Repeating this like a mantra doesn't it make any more true IMHO.
>

I'm sorry for repeating it without providing background.
Why the market would work to bring addresses into productive use, is that 
people are motivated by personal self-interest.
So a person (organization, entity) who had addresses which were dis- or 
under-used would be motivated by the money he could get by selling the 
addresses.
The buyer of addresses would be establishing, through his willingness to 
part with money for them, that he thought he could profit by having them.
If the buyer is putting these addresses to some use which is profitable to 
him, they will have moved from disuse to use through the operations, in a 
free market, of people motivated by their own self interest.
These are the basic forces at action, and I think these would be the basic 
transactions of the ip address transfer market.

Yes, a speculator or hoarder could purchase addresses and not put them to 
use, but he would only do that if he thought he could profit by selling them 
later, or profit in some other way.
A speculator or hoarder (or any other buyer) could also lose money if he is 
unable to profit from his acquisition for whatever reason.


> [and yes I think I have read all of the many posts exploring this issue]
>
> [And no, I am not against free markets per se.]
>> Prior to exhaust, without such a mechanism, I could have walked up and 
>> asked
>> for a /2.
>>
> You seem to ignore the fact for convenience that this is *almost exactly* 
> what happened at the top level allocation split of the 2^32 globally 
> unique IPv4 address space between US based entities versus the rest of the 
> world. The US has approx a /2. The rest of the World (who came rather late 
> to the party) have to make do with the rest of the 2^32 addresses.
>
> One address space does not exist without the other. We live on a global 
> Internet. ARIN allocates addresses from a proper subset of the globally 
> unique IPv4 space. ARIN members communicate with nodes located in other 
> RIR regions.
>
> If you're going to claim that a free market is the best practice for 
> handling the ARIN space upon exhaustion, then it would be completely 
> illogical to conclude anything other than that this market would also have 
> to be a global market covering all of the RIR regions and all of the IPv4 
> address space [because the 2^32 addresses of the globally unique IPv4 
> address space is a single contiguous global resource].
>
> Especially as demand for address space in APNIC has consistently grown so 
> much faster than in the US in the recent past.
>
> The real crunch with IPv4 address exhaustion is not going to hit US 
> companies internally, or even communication between ARIN members: it's 
> going to hit US companies and ARIN members who want to communicate with 
> suppliers, customers etc. located in APNIC. It's no good you having 20 
> addresses if your communication partner doesn't have any. [20:1 is the 
> current ratio of addresses per capita between the US and CN]
>
> You can't have your cake and eat it.
>
> <snip>
>
> regards,
> RayH

I understand that you feel the global address allocation is unfair, but I 
don't understand its context in this discussion about markets, except to say 
that I agree with you that the market should be global. I believe with the 
work at APNIC to remove needs requirements for transfers, coupled with the 
Global Policy Proposal to allow Inter-RIR transfer, are steps in this 
direction.
Concerning APNIC, I think the probability of unregistered Inter-RIR transfer 
is one of the reasons why they placed whois integrity over the other 
negative concerns about the implications of markets, like disaggregation. 
They decided that they had better just register the transfers as they occur, 
because they are likely to occur, in order to server the greater good of 
making sure the whois actually documented reality.
The stewards there made some basic decisions about regional demand, IPv6 
transition, and perhaps historical global allocation inequities that led 
them to approve the elimination of needs-tested transfers.

My argument is that retaining needs analysis to approve transfers is no 
longer required to ensure efficiency, and maintaining it can cause 
transactions to go off-books, to the detriment of all.

Regards,
Mike