[arin-ppml] Draft Policy2009-1: TransferPolicy (UsingtheEmergencyPDP)

Ted Mittelstaedt tedm at ipinc.net
Thu Mar 26 14:36:22 EDT 2009


> -----Original Message-----
> From: arin-ppml-bounces at arin.net 
> [mailto:arin-ppml-bounces at arin.net] On Behalf Of Matthew Kaufman
> Sent: Thursday, March 26, 2009 10:23 AM
> To: michael.dillon at bt.com
> Cc: ppml at arin.net
> Subject: Re: [arin-ppml] Draft Policy2009-1: TransferPolicy 
> (UsingtheEmergencyPDP)
> michael.dillon at bt.com wrote:
> >> the future where IP addresses are unavailable and *not even the 
> >> highest bidder* can get them is worse.
> >>     
> >
> > The science fiction list is over that way.
> >
> > Here, in the real world, we have plentiful cheap IPv6 addresses for 
> > anyone who wants them. Leading organizations such as Google have 
> > already tried these IPv6 addresses and discovered that they taste 
> > great, like a good address should.
> >
> >   
> You conveniently left out my "the IPv4 space it needs to last 
> through a transition that is starting too late to be done 
> before runout".
> IPv6 is the future. Between now and that future, there will 
> be organization that need IPv4 addresses in order to operate. 
> Some of that demand will occur after the runout.
> > What would you rather have, an intractable policy problem in how to 
> > allocate a fixed pool of scarce addresses, or a technical 
> problem in 
> > how to make IPv6 and IPv4 networks intercommunicate?
> >   
> I have yet to see a technical solution to IPv4-IPv6 
> interworking that doesn't require any IPv4 addresses during 
> the transition. And I refuse to believe that when IPv4 
> addresses are no longer available from RIRs that 
> entrepreneurs will grind to a halt and no longer fund, build, 
> and grow businesses that have a need to talk to the existing 
> IPv4 Internet.

Plenty of ISP's (like the one I work for) will have IPv4
available to be able to allow us to offer IPv6<->IPv4 gateway
products.  We will self-generate that IPv4 from our own stock
when we move our customers off IPv4 and on to IPv6.  So those
entrepreneurs who are growing businesses that need to talk
to the legacy IPv4 Internet will be able to contract with companies
like ours to be able to do it.

In any case, your assuming the entrepreneurs will shoulder the
burden of connecting their IPv6 offerings to the IPv4 Internet.
While some may, I think your going to see very rapidly the onus
will be on the IPv4 users to shoulder the burden to connect to the
IPv6 Internet.  It's those users who will become the primary
purchasers of IPv4<->IPv6 gateway services, and likely, most
of those gateway services will be offered by their own ISP
and perhaps included in their access cost.

> I have two significant issues with people who are completely 
> opposed to a transfer policy:
> 1) The idea that things have worked well so far with the 
> friendly cooperative addresses-are-nearly-free model we have 
> now and so we should keep things as-is.
> The problem is that the world we must plan for is not the 
> world we inhabit today. *There is no "as-is" after the 
> runout.* The world we must plan for is one where IPv6 
> addresses are as available as IPv4 addresses are today, but 
> IPv4 addresses are *no longer easy to come by*. And I believe 
> that because the transition is starting as late as it is and 
> going as slow as it is with an new protocol that is as poorly 
> suited to a smooth migration as IPv6 is there *will 
> undoubtedly be a continuing need for both existing and new 
> organizations to get a limited number of
> IPv4 addresses*.

Then I would suggest that those organizations with such a need
go out now and get connected and get their IPv4 while it's still

When IPv4 runs out, new orgs will not be able to go -anywhere- and
get it.  They will have to use IPv6 since they will have no choice.
So they will revamp their network to require as few IPv4 as possible
then go to a gateway provider and buy access to the IPv4 Internet.

I didn't say they would like it.  The people being told they have to
junk their TV's or get HDTV converter boxes don't like it either.  But,
in technology, the stick-in-the-mud people don't have the power to
stop change, at least, not at the current time.  They did a millennium
ago, and so we got people like Galileo being beaten into the mud.

> 2) The idea that if we don't have a transfer policy then 
> transfers won't happen.
> Existing entities which need more IPv4 space in order to 
> continue operating and/or growing, or new entities which need 
> even small amounts of IPv4 space in order to exist at all 
> (imagine, for instance, a new ISP serving a previously 
> unserved region which needs IPv4 addresses for that side of 
> their IPv6-IPv4 gateway) *will* find a way to get these 
> addresses if there is economic value in doing so. They will 
> purchase existing entities which have underutilized space, 
> they will ask entities with space to spin off subsidiaries or 
> lease address space to them, etc. 
> All of this is more legally complicated, and thus more costly 
> in lawyer time, than just being able to transfer the address 
> space. That adds an unnecessary transaction cost to solving 
> the needs of the entities which need IPv4 addresses after the 
> runout, and so that is *even worse* than a clean transfer 
> policy would be, as far as increasing the cost to the 
> acquiring entity goes.

And the downside of this is?  We WANT it to be difficult to
get IPv4 in a post-IPv4-runout world.  If it was still easy,
people would never switch until the very last drop of oil I
mean IPv4 was sucked up - then -everyone- would hit the brick

> Of course there are folks for whom the lack of a clean and 
> simple transfer policy is a win: the lawyers who will 
> orchestrate the more-complex transfers that will happen 
> anyway, and the people with existing IPv4 services who want 
> as big a barrier to new entrants as possible.

You put your biggest concern last.

Sure, the orgs involved now who have IPv4 will be at an advantage
over the newcomers.  This is EXACTLY LIKE the ISP's out there
who are legacy holders who AREN'T PAYING A CENT for their IP
numbering, while the newer ISPs who got their assignments after
ARIN are paying a lot of money.

What about all the brand-new ISP's who since they are building
networks from scratch, get to go out and buy cheap hardware
for a 10th of the cost of the expensive gear that
the older ISP's are still amortizing?  Or the newer ISP's who
started after DSL who don't support dialup AT ALL and only offer
DSL, and have much smaller support costs?


More information about the ARIN-PPML mailing list