[address-policy-wg] 2013-03 New Draft Document and Impact Analysis Published (No Need - Post-Depletion Reality Adjustment and Clean up)

* Nick Hilliard

> On 25/07/2013 15:44, Tore Anderson wrote:
>> - What kind of problems are you foreseeing?
> 
> address hoarding, 

I would actually claim that this is not difficult to pull off under
today's policy either. If you want to set up a LIR to hoard address
space, you can team up with someone who does have "need". (There are
several buyers on the Listing Service who are seeking large blocks,
perhaps a good place to start looking.) Make a proposition to them to
loan them any address space you are able to find and buy on the market
for free. With their ripe-583 forms in hand, your "hoarder LIR" has the
required need to get approval from the NCC to receive allocation transfers.

When the time comes that you need the addresses yourself, terminate the
loaning contracts and reassign/transfer to yourself as appropriate.

> leading to reduction in market liquidity,

The numbers I posted in my previous message suggest that the market is
only supplying 3-4% of actual demand (assuming last year's demand is a
good indicator of this year's). I'd say that this is not a particularly
liquid market to begin with, making it hard to see how any kind of
further marginalisation of whatever little liquidity is there now can be
a significant issue to the community when looking at the bigger picture.

> leading to profiteering.

It's perfectly possible and sanctioned by today's policy for LIRs to
make a profit from IPv4. I gave an example to Filiz earlier about how
you can, under today's policy, set up an LIR that only exists to lease
out IPv4 blocks to whatever End Users are willing to pay the most. We
were talking about the /22 allocation from the "last /8" then, but the
procedure would work equally well for transfers - simply line up End
Users and gather their ripe-582 forms, find space on the market and buy
it, assign whatever you bought to whoever is first in line (or willing
to pay the most). Lather, rinse, repeat.

When a certain allocation has been in your possession for 24 months, you
may also choose to transfer it away, or just keep leasing it out of
course. Do whatever maximises your profits. You're not in any way
breaching current policy, after all, you're essentially just operating
an LIR.

I'm not saying this is "good" or this is "bad", all I'm saying this can
be done today, it's not something new that 2013-03 brings to the table.

>> - What kind of control mechanisms exist in today's policy, that are
>> removed by 2013-03, will forestall these problems?
> 
> historically, two things: the presence of policies which make it difficult
> to transfer large amount of address space from many different sources,

Could you point out which part/passage of the policy you're referring to
here?

I am under the impression that if you have approval from the NCC for,
say, a /8, there's no language in the current policy that makes it more
difficult to transfer 16384 /22s from 16384 different sources, compared
to transferring 1 /8 from 1 source.

[I have absolutely no problems seeing that performing lots of small
purchases is wildly more difficult than performing one large one, but
for reasons that has nothing to do with any current policy language.
Locating all the sellers and striking a deal with all of them, for
starters.]

> secondly, the fact that IP addresses had little market value due to the
> fact they were effectively being dispensed for $notmuch by the rirs.  I.e.
> lots of effort for little return.

But - this already changed when the RIPE NCC ran out of space and
implemented the last /8 policy last September. I don't see what change
is brought by 2013-03 in this regard? Could you point to the
changed/removed policy language in question?

>> s/24-month reallocation freeze/requirement to demonstrate need/
>>
>> Does this argument work equally well?
> 
> it's the same argument, which means that we need to understand that it is a
> bad idea to put any faith in the 24-month reallocation freeze rule as a
> means of regulating the market, or to believe that the rule will stay there
> as a long term fixture.

So just to make sure I have not misunderstood you here, you are also
accepting the following argument to be equally valid (or invalid) as the
one you wrote above?

"we need to understand that it is a bad idea to put any faith in the
requirement to demonstrate need rule as a means of regulating the
market, or to believe that the rule will stay there as a long term fixture."

Tore