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

[Owen DeLong]

On May 11, 2011, at 7:41 PM, Mike Burns wrote:

> 
> Hi Owen
>> 
>>> I have personally seen many asset sale agreements which included legacy IP addresses which were completed without notifying ARIN, as there is still no >>requirement for legacy holders to do that. I have seen asset sale agreements which include ARIN accounts and passwords among the listed assets. The >>addresses change control, but whois still shows the original registrant.
> 
>> There is no reason that ARIN, upon conducting a section 12 review and finding that the original holder was
>> no longer using the resources and/or was defunct could not reclaim those resources, so, the recipient in that
>> case is taking a rather large risk.
> 
> Wrong, may I quote NRPM 12.8?
> This policy does not create any additional authority for ARIN to revoke legacy address space.
> 
Correct... no _ADDITIONAL_ authority. ARIN has always maintained that they have the ability
to reclaim abandoned legacy space. If the original resource holder is no longer using the space
and/or is defunct, ARIN has the right to reclaim the space. If the successor organization chooses
to complete an 8.2 transfer and sign an ARIN standard RSA and update their whois, that's fine.
However, if they do not, then, I see no reason for ARIN to leave bad data in whois to allow them
to continue to hijack the space in a manner contrary to the policies set by the community.

> 
>> I do not believe ARIN will view the asset sale agreement as legitimizing
>> the transfer and the community developed policies support reclamation of resources from defunct legacy
>> holders.
> 
> I never said they were defunct, they may have simply been acquired.
> 

ARIN has a moral, if not legal (the legal status is somewhat unknown) to continue to provide
services to the original legacy registrant of certain resources. That obligation is non-transferable.
In the case of an acquisition, the acquiring party has the right to keep the resources to the extent
they are still justified under current policy by completing a transfer according to section 8.2 and
signing the standard RSA.

To the best of my knowledge, ARIN has no such obligation to a successor organization that
chooses not to operate within policy and complete the transfer under section 8.2

Certainly, ARIN has no obligation whatsoever to a third party that is under the mistaken impression
that they were able to acquire the addresses independent of the underlying network and purpose
for which the addresses were originally issued.

>>> When it comes time to route the addresses, if the network operator questions the situation, I have seen them accept the asset sales agreements as acceptable >>proof of routing authority. And the addresses allocated to entity A are now in control of entity B, with bogus whois data. This is the kind of eventuality which I >>believe motivated the APNIC community to place the stewardship role of uniqueness above the stewardship role of needs-based transfers. Obviously I am >>asserting these things without documentary proof.
>> 
> 
>> Again, can you cite an independently verifiable example? If not, then, I still believe this is a straw man argument.
> 
>> How many such agreements have you seen? (I tend to think that it represents a relatively small fraction of
>> the address space and will likely get resolved through the current contact validation process for the most
>> part.)
> 
> Owen, if it were just Mike Burns saying this, would APNIC have changed their policies?

APNIC did not "change" their policies any more than ARIN changed their policies.

APNIC enacted a transfer policy. ARIN enacted a transfer policy.

In the development of those policies, a difference emerged between the policy developed
in APNIC and the policy developed in ARIN.

In APNIC, the transfer policy does not require justified need.
In ARIN, the transfer policy absolutely requires justified need.

Both were new policies developed by the two regions, independently, but, at approximately the
same time. I believe that the policy in APNIC is largely the result of very strong and vocal lobbying
by Geoff Huston and a couple of other individuals that routinely take a rather dominant role in
the policy development process there. Even with such strong and vocal support from community
leaders in the APNIC region, the debate was long and the outcome was not entirely certain
until the matter was finally decided. If you have never been to an APNIC meeting to observe
how they determine consensus around a proposal, then, I suspect you have no frame of reference
to properly evaluate just how little the fact that APNIC chose to abandon needs-basis actually
means in terms of using it as evidence to support such a  claim.

> Yes, the agreements I have seen represent a small fraction of the address space.
> No, I can not cite an independently verifiable example, I hoped by now you would have thought higher of my intelligence.
> 

This has nothing to do with your intelligence or my opinion of it. Without an independently
verifiable example, it remains assertion and conjecture rather than proof or fact.

You are asserting that these things have happened. I am questioning the validity of that claim.

You are asserting that they are happening to such an extent that the integrity of whois is in
jeopardy if that continues. I am saying that for the community to properly evaluate such a
claim, we would need visibility into the nature and scope of such transactions.

>>> While I don't think that's the entire argument or even a particularly accurate framing of that portion of the
>>> argument, I would say that the history of free markets does give one plenty to fear. Especially when you
>>> consider that history in situations of truly finite (even for a short time) resources. (Tulips anyone?)
>> 
>> > You seem to consider that needs-based allocations were some kind of > social agreement preventing malfeasance of the wealthy and protecting > the little guy.
> 
> Indeed, i do.
> 
>>> In reality, it was the least  rulemaking possible in an era of free-pool allocations.
> 
>> No, we could easily have handed addresses out for the asking without any regulation whatsoever.
>> We, as a community chose not to. We made that choice for good reasons. I believe those reasons
>> remain and may even be enhanced by runout.
> 
> C'mon Owen! Who could possibly believe what you are saying here? Of course there were good reasons for not handing them out without any regulation whatsoever. I have expressed over and over that nobody has ever supported that notion, to my knowledge. Needs assessment was a noble and valid means of stewarding addresses into productive use.
> 
First you question who could believe what I am saying, then you agree with it. I am confused.

>>> There really was no other way to distribute scarce resources with no price unless the allocations were limited by need.
> 
>> Not at all true. They could have been simply handed out to whoever asked first. I agree such an
>> approach would have been ridiculous. The difference between us is that I believe even when you
>> include dollars in the process, such an approach remains ridiculous because it transfers the
>> address regulation from community based stewardship to a system where the only factor
>> determining resource allocation is the accumulation of capital.
> 
> Owen, like it or not, dollars are going to be involved in the process of address transfers.

I never said they would not. Indeed, I am not even opposed to that. However, I am opposed to
allowing them to be the sole and only mechanism of regulation.

> That is not at issue here.

Correct, to a certain extent. (See above).

> Thanks for agreeing that your proposed alternate mechanism for allocation from the free pool is ridiculous, it should be a short step from there towards acknowledging that the needs requirement was the lightest touch of a proper steward allocating from the free pool.
> 

I think you are entirely missing my point. I am not sure whether you are doing so deliberately or otherwise.

My point is that removing needs-basis from allocation policy makes as little sense with money involved
as it would have without money involved. There is nothing to align the money with the community interest
or good stewardship.


>>> And nobody really debates that, even an outlier like me.
> 
>> Interesting... I will actually debate that. I agree the alternatives were absurd. However, to claim that
>> they did not exist would be as logical as my claiming that an unregulated market is not an option.
>> It's an equally detrimental option, but, it's an option.
> 
> Would you at least admit that there was no other option if we wished to be proper stewards, both ensuring productive use and making the minimum of rules to allow for that?
> 

There is no other option today if we wish to remain good stewards. The option of abandoning needs
basis is equally absurd now as it was then and for the exact same reasons.

Yes, we now have to allow money to become an additional factor. It remains to be seen exactly
what effects (good and bad) that will have in this environment. However, I greatly prefer the option
of adding money as an additional factor vs. replacing all current regulation with money as the sole and
only regulation. I think most systems administrators would agree that making a single change and
evaluating the results is a more reliable approach to a stable system than making multiple simultaneous
changes.

>>> That cause goes away with the free pool, and the imperative against more rules than necessary dictates we lift those needs-based transfer rules.
> 
>> Indeed, we do disagree. I believe that the cause for needs-basis was the idea that the community
>> preferred not to grant resources to entities that did not need them to the exclusion of entities that
>> do.
> 
> No, I think we agree. At least I agree with your last statement, if we limit it to the free pool, which was the context.
> 

I don't see any reason that statement doesn't or shouldn't apply equally to the movement or retention of addresses
after the free pool is exhausted, either.

> 
>> I see no way in which bringing money into the picture changes that reality or that community ideal.
> 
> Owen, you are conflating issues. Money is coming into the picture regardless of needs requirements.
> 

No, you are confused.

My statement is that merely bringing money in does not change the picture and does not change the
reality of that community ideal. I accept and understand that we are bringing money into the picture.
Where I am disagreeing with you is the belief that that provides a good reason to abandon the more
rational approach at the same time.

I believe that money and needs-basis can coexist and result in a better overall outcome for the
community than the one achievable through money alone.

> 
>>> We have to understand the cause of needs requirements. It wasn't some egalitarian ethos, it was an obvious and fair mechanism for placing some limit on >>address allocations from a free pool of limited size.
> 
>> And it still is. The source pool for IPv4 address transfers is not unlimited in size, either.
> 
> No, it is not unlimited, but now we have price to ensure productive use, something we did not have when allocating from the free pool, so we chose the least intrusive mechanism to ensure that at the time, which was needs requirements. Now we have no free pool and the mechanisms of supply and demand to ensure productive use.
> I agree that needs requirements be continued on free pool allocations.
> 

Price does NOT ensure productive use. No matter how many times you repeat this absurd
fallacy, it remains an absurd fallacy. Since price will not ensure productive use, abandoning
needs basis (which has proven reasonable effective at doing so) does not make sense.

When you build your argument on a flawed premise, your conclusions tend to amplify that
flaw. Price ensures that the participants in the market will act in their own financial interest
with minimal, if any, consideration for the impact on the other members of the community.
Indeed, with price as the only control, anti-competitive manipulations become not only possible,
but, very likely.

> 
> 
>>> We didn't impose max limits per allocant, we didn't impose progressive fees that made larger blocks proportionally more expensive, we didn't create rules to >favor corporate diversity, we didn't limit distributions per country, we didn't require an income statement of recipients so that we could judge whom to allocate to.
>> 
>> No, we chose not to allow greedy entities to absorb more space than they needed
>> to the exclusion of others. Your claim is that adding money to the situation somehow
>> removes the need to do so. My argument and belief is that it does not.
> 
> Again, Owen, this is not about money. Money is going to be part of all transfers.
> This is about maintaing a needs requirement.
> 

Try reading what I actually wrote instead of responding to the internal monologue.

I am not saying that money isn't coming into the situation. I'm saying that just because
it does, doesn't remove the need to prevent greedy entities from absorbing more space
than they need to the exclusion of others.

You are saying that bringing money in fundamentally changes the need for regulation.
I am saying that it does not. You cannot argue that bringing money in makes all the
difference on one side and then claim that the rest of the discussion isn't about money.
The question is how money will affect the proper stewardship of addresses. I believe it
makes little difference, and, to some extent makes it more difficult to maintain proper
stewardship because prior to involving money, people were more likely to do the
right thing just because it was the right thing to do. When you add financial incentives
which may or may not be aligned with acting in the community interest, that changes
things, but, not for the better. Certainly not such that removing prior safeguards makes
any sense whatsoever.

>>> I have searched long and hard for a historical analog to this situation. The best I could find was a policy of the US after the Revolutionary War, which allocated >>property to veterans of that war for free. You had to qualify by being a veteran, the total property available for allocation was limited in size, and the property >>had value.
> 
>> Not an accurate analogue. There was also other land available through a variety
>> of other land grants, purchases, and even in some cases, just being the first to
>> arrive somewhere and stake a claim.
> 
>> Indeed, since there was no "justified need" basis for land allocation at any point
>> prior, the fact that things continued without it on a relatively even keel is kind
>> of irrelevant.
> 
> The justification was that you had to be a veteran who fought in the Revolutionary War.
> 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.

> 
>> In the historical case, there were no limits on resale of the allocated land, there were aggregators, there were speculators, and things progressed normally during ?>the allocation time period, and the time period after the land was fully allocated.
> 
>> But there was nothing like needs-basis in the initial allocation and there were
>> many other sources of land as well. This simply isn't the case with IPv4.
> 
> There was something like needs-basis. You needed to demonstrate that you were a veteran.
> Land was not handed out to just anybody. (Remember, this land had value.)
> Sure you could move somewhere else and purchase land, like you can choose to deploy IPv6 or CGN.
> 
Land was being handed out to just anybody. Maybe not that particular land, but, other land.
In fact, there were efforts to recruit immigrants from foreign countries based on the promise of
land grants.

>>> Some people intended to live and farm the land, others intended to sell their allocations. And American law allowed a free market in which these men could decide.
>> 
>>> Rather than judge the benefits of free markets by exceptions like manias and successful market cornering, both rare and shortlived events, why don't we judge >>free markets like our American ancestors did, even if we're Canadian, eh!
> 
>> Completely unregulated free markets have never functioned. All markets
>> degrade to the point of requiring the government (or some industry
>> organization) to step in and mitigate the issues they create. This is
>> not the exception, it is the reality we have seen time and again.
> 
>> Preserving needs basis is merely a rational and good form of regulating
>> the IPv4 transfer market to the benefit of the community.
> 
> What is the great benefit that outweighs the danger to Whois of unbooked transfers and the benefits of incentivizing legacy resources to come under RSA?

Since your intent is to gut the RSA to the point of rendering it nearly meaningless, I fail to see much benefit remaining
in getting legacy resources under said RSA. Indeed, removing the most meaningful provisions of the RSA from those
currently subject to its provisions would, IMHO, do much more harm than any amount of unbooked transfers will
likely ever do.

> Who says my proposal would result in completely unregulated markets? Every transferee would sign an RSA and be subject to ARIN policy.

Except that you want to render ARIN policy and the RSA virtually moot in the process.

> 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.

> If the policy proposal doesn't fly, it will be to the highest bidder who can show need.

More importantly it will be to the highest bidder that can show need and only to the extent
of his need, leaving the remaining addresses for some other bidder.

> Do you think that we should take more steps to protect the little guy in this event?

I don't know. I'd like to see how things develop with the current policy and re-evaluate once
we have some experience. As it stands, I do not believe there is community will to do so.

> Maybe a price cap? An address czar?

Believe it or not, in general, I am opposed to artificial price caps. I hope that by preserving
needs basis and removing the opportunities for speculators and other wholesalers
attempting to artificially increase prices, we will be able to avoid any need for them.

Since I'm not sure what you mean by an address tsar, I have no idea.

Owen