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

[Mike Burns]

Hi John, thanks for the feedback, responses inline.

>
>> 1. Policy Proposal Name: IPv4 Transfer Policy Change to Keep Whois 
>> Accurate
>
> I note with interest the explanation that the Policy Proposal Name is 
> intended to be taken with tongue firmly in cheek. I'm glad that's cleared 
> up as I was fixin to expostulate,
>

I will correct that when I send it to policy at arin.net for official 
distribution.


>>
>> 10.    ARIN will not use utilization as a measure of policy compliance 
>> for addresses transferred under 8.3.
>
> because I was confused as to why a policy whose major (only?) proposed 
> effect is to remove utilization devotes most of its wording to WHOIS 
> accuracy. Who wouldn't want that?  But I understand now that it is 
> intended only as a witticism. I will therefore attempt to speak only to 
> the non-WHOIS part of the proposal.
>

No, there are two major effects to my proposal, one related to removing 
needs-requirements from transfers, the other allowing existing RSA holders 
to sell unused IPv4 addresses without fear of a review/return request if 
they have had them for more than a year.
I copied the proposal largely from the APNIC proposal which passed there, 
and I believe that the Whois accuracy argument is the most persuasive, the 
name is a witticism, not the rationale.
I fully believe that if Microsoft chose not to sign an LRSA with ARIN, or if 
ARIN had never become aware of the deal, MS could have legally consummated 
the deal with Nortel, and that somebody would have routed the addresses for 
Microsoft.
The net effect would have been Whois data for those blocks showing 
allocations to the original and long defunct entities, instead of pointing 
accurately to Microsoft.



>>
>> 8. Rationale:
>
> <snip>
>
>>
>> The underlying proposition behind this policy proposal is that the
>> registry of IPv4 addresses operated by ARIN is of general utility and
>> value only while it accurately describes the current state of address
>> distribution. If a class of address movement transactions are excluded
>> from being entered in the registry, then the registry will have
>> decreasing value to the broader community, and the integrity of the
>> network itself is thereby compromised.  This proposal's central aim is
>> to ensure the continuing utility and value of the ARIN address
>> registry by allowing the registry to record transactions where IPv4
>> addresses are transfered between ARIN account holders.
>
> But where is utilization in this underlying proposition?

As far as I can tell, the word utilization in the paragraph above refers to 
the utilization of the whois registry, not address space.
I have argued that when things are valuable, especially if they are valuable 
for a short time, they will be sold if not needed.
To me, this will naturally increase utilization, I point to the billion or 
so unrouted but allocated addresses as some evidence that there is more 
efficiency to be wrung from these allocations.
Yes, I understand that some allocated but unrouted addresses are used 
internally, my perception is that this is a subset of unadvertised but 
allocated space, much of which is simply wasted.
My argument about needs requirement is that it is an artificial contrivance 
designed to constrain free pool allocations whose purpose has been subsumed 
by the mechanism of price in the transfer market.
And that maintaining that market artificiality will result in some 
transactions being performed privately, and because the addresses will still 
work for their intended purpose whether or not ARIN is notified, and because 
there is no legal venue to challenge the transaction if it involves legacy 
space, those transactions will not be reflected properly in Whois.

>
> Having observed the preparatory discussions that led to this proposal and 
> subsequent comments to it, I recall massive dialogue on the MS/NN 
> transaction. To integrate that subject with this proposal, I must go to 
> comments to the proposal for guidance. In the authors response of 11 May 
> 2011 17:12:57 to Tom Vest regarding "current legal realities" it is 
> stated: "4. Finally, the requirement at issue here, a needs analysis had 
> to be completed by ARIN, which magically showed that MS qualified for 
> exactly the amount already bid for and negotiated the sale of."
>
> And this paragraph
>
> "If they had needed fewer addresses, they could have bid for less than the 
> full pool. If they needed more addresses, they should have received an 
> additional ARIN allocation. Paraphrasing Goldilocks, the random allocation 
> of addresses to long-ago Nortel acquisitions was "just right."
>
> which I interpret as expressing disbelief. I could be wrong here.
>

No, you are right, I was expressing disbelief. I believe that all the other 
issues with the transfer as negotiated prior to ARIN's knowledge of the deal 
were ably finessed so as to fit the deal into the existing policies, via 
some ex-post-facto 8.2 transfers, stretching "RSA" to include "LRSA", 
syntactically applying "single aggregate" to need and not to receipt of 
addresses, and considering the "issue to ARIN"  language to be a mere 
procedurality.

To my mind, the most difficult effort in shoehorning the transaction into 
existing policy was the needs requirement. I believe that requirement 
contorted ARIN into knots, and at least put ARIN into the potentially 
corruptible position of  deciding the fate of a $7.5 million dollar 
transaction.

I believe that had my proposed policy been in effect, the deal would have 
been simply booked into Whois, the market would have put addresses right 
where they were needed (according to ARIN), unused addresses would have come 
off the shelf, and those addresses would be put under the same RSA that 
every other RSA holder is under, and questions about ARIN trust and 
potential corruptability would be absent.

>
> 2) It looks to me like we have a fundamental logic flaw here. This 
> proposal will fix a problem that does not exist! From my perspective, a 
> needs assessment was properly conducted in the Microsoft/Nortel bankruptcy 
> affair and ARIN policy was followed. John Curran says so, author says no. 
> Which brings us to the realm of belief and he said, he said. Who do I 
> believe, John or author? Well, John has been exposed to the facts of the 
> matter and cannot talk about it, except obliquely. The rest of us have 
> only our interpretations of public documents. John has a fiduciary 
> responsibility here, so real penalties might apply if he prevaricates. 
> Author can pretty much say what he likes, the worst we could do is get 
> quite cross. John has several decades of credibility in public policy and 
> networking. So might the author, but I can't say. But here's the deal, IT 
> DOESN'T MATTER. We will never know and can't know.

We may never know, but have you considered an NDA with a time limit?
Or WikiLeaks?

 So may we please
> dispense with the APNIC did it thing? It's a fallacy anyway: 
> http://en.wikipedia.org/wiki/Argumentum_ad_populum

> respectfully
>
> John Springer
>

I will make a deal. If aspersions of poor stewardship can be done away with, 
I will do away with the APNIC references.
It doesn't advance the argument and doesn't advance our relationship with 
the APNIC steward community.
It is easy to declare one person a poor steward, it is less believable when 
another steward community is being included in the smear.
I have already requested that posters stop equating my proposal to remove 
needs tests from transfers to phrases like "jettisoning stewardship".
We are arguing over proper stewardship, as stewards, and I have tried to 
refrain from accusing those who oppose my policy of being poor stewards of 
Whois.

Regards,
Mike