[ppml] Policy Proposal: Resource Reclamation Incentives

[Ted Mittelstaedt]

>From: ppml-bounces at arin.net [mailto:ppml-bounces at arin.net]On Behalf Of Owen
DeLong

>>On Jul 2, 2007, at 11:26 AM, <michael.dillon at bt.com> wrote:

Owen, it's bad form to post HTML mail pleast stop it.  We all aren't
running non-complaint mail clients and it makes quoting difficult.

>> but the legacy holders are closer to
>>being here illegally than those who sign the RSA and pay their
>>membership fees.

>No. They are not. The industry has, generally collectively agreed that
>legacy holders are grandfathered under a different set of rules. The fact
>that you don't like this collective decision is another matter.

The industry has grandfathered the legacy holders into the system
because it is to the industries benefit, NOT to the legacy holders
benefit.

Because the Internet is IPv4 now, we need to know who the legacy holders
are.  Thus it benefits us to keep an eye on them.

Once we switch to IPv6 there will be less and less incentive for the
rest of us who have switched to IPv6 to worry about IPv4 numbered sites
that don't want to play by the rules.  Thus there will be a huge incentive
to boot the non-paying legacy holders out the door.  I would push for ALL
records for commercial legacy holders to be stripped from the RIR's after
IPv4 exhaustion, and for non-profits to be stripped 2 years later,
unless that is they sign RSA's and start paying money to the RIR's.

Those IPv4 addresses can then be available for reassignment.  If I'm an
ISP that desperately needs IPv4 addresses post-IPv4 exhaustion, and ARIN
tells me "we don't got any clean addresses for you, but we will assign
you addresses that a legacy holder is currently using and has no right to
use on a temporary basis" then you bet your ass that I will start
advertising
those addresses. The legacy holder can then go bitch to his interconnects,
I will sit tight
with my allocations, and all of the intermediate networks will see that I
have a right to those numbers (because I'm in the whois) and the legacy
does not (since they are in nothing) and who do you think is going to "win"

I may not be able to use those new IPv4 addresses for a year, perhaps, but
during that year things will be extremely difficult for the lgacy holder,
their users will be screaming at them because of being unable to get to
various websites and so on, and that will provide incentive enough for
the lgacy holder to sign an RSA and start paying.  Or to vacate the
addresses.

>>They could be deported, i.e. the legacy resources could be taken away
>>from them by suing them in court.

>Well... According to ARIN's lawyer, we probably wouldn't win on that
>one, so, I'm not inclined to believe your statement over that of Steve
>Ryan.

I don't see the need to sue anyone.  If the lgacy holders who refuse to
give up their IPv4 allocations and switch to IPv6 want to go off and form
their own little IPv4 Internet with their own RIR then more power to them.
It will be the right of any IPv6 site to block IPv4 access from the rogues
that don't want to play fair.

Extremely large legacy organizations are not going to give up access to
customers on IPv6 networks, that serves as enough incentive for them to
switch over.

>They aren't subject to ARIN rules. They have no contractual relationship
>with ARIN and there is no legal basis for claiming that they should be
>subject to ARIN rules. ARIN has no force of law other than the contractual
>relationships they have with the recipients of ARIN resources.

>So far, nobody seems to be boycotting Harvard or MIT because of their
>legacy address space. I don't think such a thing is likely in the future.
>I don't know of any organization who is losing business because of their
>possession of legacy addresses. Do you?

That isn't the issue.  The issue is that Harvard and MIT have students
that will want to go to websites that will eventually switch over to IPv6,
(because IPv4 will not be available) and if those sites are boycotting
legacy holders, then Harvard's own students will start agitating for Harvard
to fix things.

And all it takes is one angry man with the force of law behind him.  If
ARIN withdraws sponsorship of Harvards' ARIN whois records, and assigns
the IP subnets that, lets' say, Harvard's nameservers are on to Mr Michael
Dillion, just how long do you think that Harvard will have a viable
Internet connection?

Harvard and MIT may be big and powerful compared to you or me, but they
are nothing compared to the rest of the Internet.

>Secondly, I think operationally, such actions against the larger holders
>of legacy addresses (i.e. the ones that matter in terms of this policy)
>would be unlikely because, generally, ISPs don't want to piss-off
>large clients.

What is your definition of an ISP?  I hae read about a single individual
in the past who gamed the system and now has something like a /18
and has no more ISP to his name than Burger King does.

>Court action has been deemed unlikely to succeed by
>someone I am convinced knows way more about it than you do, so,
>I think you're wrong on that as well.

Ah, yes, I'd love to see this one:

Harvard Plaintiff:

  "Your Honor, Defendant is wilfully disrupting our Internet service to
thousands
of students and costing us millions of dollars by using the IP addresses
that we
were assigned"

Judge:

  "Assigned by who?"

Plaintiff:

  "well they were assigned by these dead guys out of a spiral notebook
sometime
about 40 years ago"

Judge:

  "Is this how IP addresses are assigned in this industry?"

Plantiff:

  "Uh, well, no not really"

Judge:

  "Baliff, please throw these people out of the court, suit has no grounds
to
be filed"


>>In fact, given the unlikeliness of an organization going through the
>>pain of renumbering to be a good network citizen, I suspect that this
>>policy was introduced as an attempt to weaken ARIN's case-law position.

Interesting.  Do you apply such a litmus test to all ARIN proposals,
Michael?

>You can suspect all you want, but, I can tell you that I am pretty sure
>I know better than you the intent of the introduction of this policy.
>The intent is to remove some of the barriers to address space reclamation
>and to encourage legacy holders to begin using IPv6 and join
>the ARIN community and process.

I completely disagree with the idea of giving away more free stuff to legacy
holders.  I also completely disagree with the idea that the legacy holders
should get away without signing an RSA.

ARIN and the RIR's have a big giant stick they can use with the legacy
holders,
the threat of withdrawing whois records for legacy holders.

I believe the first step needs to be to tell the legacy holders that by the
time IPv4 runout occurs they MUST HAVE SIGNED an RSA that committs them to
paying ARIN for IP assignment.

If you want to carrot and stick them, then tell the legacy holders that
runout is planned for year 2012, and for every year they delay signing an
RSA they will lose a year of fee deferrment.

In other words, if they sign right now, 5 years in advance of planned
runout, they will get 5 years of fee exemptions POST runout.  (ie: fees will
be exempted until 2017)

If they sign next year, 4 years in advance of planned runout, they get
only 4 years of fee exemption post runout, (ie: no fees until 2016)

If they sign 3 years they only exempt until 2015.  And so on.

If they do not sign at all then on planned runout, 2012, their records
will be struck from WHOIS and their IP numbers will be allocated to
new requestors post actual runout.

The only way that a controlled transition will EVER happen from IPv4 to
IPv6 is to get EVERY holder onto RSA's then start jacking up fees to
retain IPv6, with exemptions to organizations that have dual-stacked.

As the years go by the fees for single-stacking on IPv4 will get higher and
higher and the discounts for dual-stacking on IPv4 and IPv6 will get
higher, and IPv6-only sites will get the steepest discounts.

I simply do not believe an orderly transition can happen if a large
percentage
of IPv4 holders are not under RSA.

Ted