[arin-ppml] Policy Proposal: IPv4 Recovery Fund

[Leo Bicknell]

In a message written on Sun, Nov 23, 2008 at 11:11:14AM -0500, Milton L Mueller wrote:
> Don't agree - you haven't reached a failure, you've reached a situation
> that prevails for 90% of all material resources in society. Anyway, I
> suspect we are near that situation now. (It's a guess, of course)

I actually see three distinct phases that will occur, no matter
what transfer policy is or is not passed.  What the different
policies do is change the timing of the transitions between phases,
primarily.

Phase 1: The current state, there is a free pool.

Phase 2: There is no more free pool, but enough people are willing to
         give up space at a price someone else is willing to pay.
         Demand continues to be met, probably with slowly increasing prices.

Phase 3: No one is willing to part for space without excessive sums of 
         money and no buyer can afford those sums.  Demand is no longer
         met.

If we ignore the black market in the current situation the white
market effectively goes from Phase 1 to Phase 3 on run out.  All
of these transition policies create Phase 2, and make it exist for
some period of time.  It would seem to prudent thing is to make
phase 2 as cost effective to the industry as a whole as possible,
and to make it last as long as possible.

At least, that is the basis on which I evaluate all of the transfer
proposals.

> What you may not see is that supply and demand is always relative to
> price and never absolute. That is, the concept of "wanting space"
> depends on "what do I have to give up to get the space" just as the
> concept of giving up space is contingent on "what benefits do I get (or
> forego) by giving it up"? if you tell people that the price of a /8 is
> $30, I guarantee you you will have 5 times more applicants for it than
> there are /8s. 

With all due respect, I think you're taking the ivory tower economics
professor view of the world.  You are 100% correct, but I fear
you're complicating the economics in a way that's not helping the
evaluation of the proposal.

Price and demand are always elastic.  If there was one loaf of bread
left in the US I have no doubt that Bill Gates or Warren Buffet
would end up spending billions to get it.  However, the statement
"the people have no bread" is still accurate.

There is some threshold at which I think we can all agree "there
are no more IP addresses available" is a reasonable generalization.
We might argue over if it is when 50% of the people in line can't
get any, or 99%, or 99.99999%.  However, much like bread and starving
people, my feeling is the threshold is lower than you may think.
No ruler survives 30% of his subjects starving; if the supply of IP
addresses leaves 30% of the ISP's wanting I suspect most ISP's will
say "we are out"; even though there is some long tail of the market
still to play out.

> > That is where these transfer proposals are doing any good.  What
> > people care most about when that is happening is price and safety.
> > They wants systems that insure they pay as little as possible, and
> 
> I don't think regulating the price of transfers is a legitimate concern
> of ARIN

I agree, which is why I have focused on transparency rather than
regulation.  E-Bay is an interesting example; in most collectable
markets E-Bay did two things; drove the costs down for most things
and drove the costs up for truly rare things.  It did this not via
regulation, but via transparency.

When a buyer can see there are 10,000 people offering the same
baseball card prices fall.  No longer are they making an uninformed
decision about the price sitting in their local shop with a shop
keeper feeding them a line of BS about how good of an investment
they are getting.  But also, when someone has the actual one of a
kind misprinted card they can put it in front of more buyers, driving
up the price.

That's why my focus is on price transparency.  Early in phase 2
(above) this will drive the price down, and help prevent people
from being swindled.

Neither 2008-2 nor 2008-6 had any way to enforce price transparency.
The value of all transactions could be NDA'ed to the two parties
involved.  This is the opposite, the value of all transactions is
published; no regulation, just transparency.

> I have always had trouble with 2008-6, not for the reasons you do, but
> because it was so obviously inferior to 2008-2. Since it is
> incomprehensible to me why 2008-2 was rejected, it is hard for me to
> know what to "support" in this context or what difference supporting
> anything will make.

Here's my own $0.02.  People are worried about crooks, and speculators,
and hoarding, and all the potential problems of a market.  2008-2
had many provisions to regulate the market to help prevent those
problems.  However, ARIN had no real mechanism to enforce those
regulations which means they could be freely ignored.

ARIN is not a government agency.  The only way ARIN can effectively
enforce anything is via contract law.  If a party who has no contract
with ARIN (a legacy holder) sells to another party with no contract
with ARIN (a new entrant perhaps) ARIN has no resource in that
transaction.  The best thing ARIN could hope is that it's members
end up enforcing that not being part of the community was bad, by
refusing to route the prefix since it is not in whois.

So we end up arguing over enforcement mechanisms (revoking space,
charging fines, flagging in whois) which everyone releases in
effective and thus support for the proposal erodes.

In 2008-6 we have the opposite.  The authors looked and thought,
ARIN can't enforce any of this, so why write down a bunch of rules
that we can't enforce.  While this "cleans up the policy", as there
is no need to argue over enforcement that will never be effective,
it completely leaves the problems of fraud, speculators, crooks,
brokers, etc.  The community seems unwilling to allow a free for
all to take place, so I don't see a lot of support for this proposal
as well.

That's why I came back with this new approach.  Can we find a way
for the industry to bring some limited self-regulation to the table;
enough to keep the really bad things from happening but not so much
that there is a burden.  Since ARIN has only contracts to enforce
regulation a fundamental facet of the design seems to be getting
contracts in place.  We saw this in the 2008-2 argument with people
wanting the person giving away space to put it under RSA/LRSA prior
to giving it away.

Thus a fundamental principal of this proposal is breaking the problem
into two, two party transactions under a contract (note, not necessarily
the RSA/LRSA):

1) Party giving up space and ARIN.

2) ARIN and the party receiving space.

This gives ARIN a solid contractual leg for enforcing some limited
amount of regulation.  Even better, ARIN already does #2 every day;
the contractual and business process aspects of it are done.  As such
the overhead is extremely small.

It also feels a lot like the current process.  If you need space you
keep submitting forms to ARIN, like always.  Almost nothing changes.
There's nothing to explain.  There should be no bumpy transition to some
other scheme.

The question then becomes what should ARIN regulate?  I'm not a fan of
regulation, so I'd like to keep this simple.  There are several obvious
classes of badness:

Speculators: Need-based pre-qualification is one way to keep them out.
Fraudsers:   They are selling to ARIN now, which has the records and
             experience to vet that they do not own what they are trying
             to sell, as compared to a random receiver who would not.
Crooks:      People who might take your money and then not given you 
             the goods.  ARIN takes on the risk, not the end users,
             but they are also better prepared to deal with it because,
             like the fraudsters they can figure it out before doing
             a deal.
Hoarders:    Won't be able to pass needs-based qualifications.

So rather than regulating these out of business, they simply can't
pass muster in this basic scheme and go away.  The "monkey in the
middle" takes significant risk off the recipients; which is why it
is good.  This is also the beauty of the system, rather than passing
complex regulation (see all the provisions of 2008-2) most of these
problems go away with no specific regulation due to the structure
of the transactions.

If you agree with the scheme so far, there's only one problem left
and it is the problem you've zeroed in on.  Pricing.  What scheme
do we use to set the price?  In terms of transparency and predictability
I think having the transactions known is good; but quite frankly
E-Bay could do that just as well.  It's not a feature ARIN is
necessarily uniquely qualified for; however by centralizing it I
think there is some good of having a single place to look.

But it's really all about the "auction format".  High bidder, dutch
auction, does first come first served play a role, etc.  That's the
only real variable in this proposal.  I suppose you could also
consider if it is an auction in both transactions; in this proposal
I have ARIN set the asking price for blocks based on bids; which
seems way simpler than having some sort of bidding on the sales
price on that end of the transaction.

But lastly, to your point about 2008-2 and 2008-6, and even this
propsoal.  While economics and regulation play into this, it's also
a sandbox built by the people who will use it.  ARIN is comprised
of the people who will need to buy and sell space, of the people
who will have to deal with these processes, and the people who will
have to make future policy.  I don't see a lot of people at ARIN
members wild about trolling e-bay to get their next block; they
want a more preditable system even if that means it may not be quite
as economically or regulatorly efficient.

I see some wins to this proposal, the question now is, how does the
rest of the community feel?

-- 
       Leo Bicknell - bicknell at ufp.org - CCIE 3440
        PGP keys at http://www.ufp.org/~bicknell/