[arin-ppml] The Library Book Approach to IPv4 Scarcity

[Ted Mittelstaedt]

> -----Original Message-----
> From: John Santos [mailto:JOHN at egh.com] 
> Sent: Wednesday, November 12, 2008 4:03 PM
> To: Ted Mittelstaedt
> Cc: 'Chris Grundemann'; ppml at arin.net
> Subject: Re: [arin-ppml] The Library Book Approach to IPv4 Scarcity
> 
> 
> On Wed, 12 Nov 2008, Ted Mittelstaedt wrote:
> 
> [I think Chris Grundemann wrote the "> >" stuff]
> 
> > > entry).  In other words, we must maintain open access to the
> > > IPv4 Internet until the IPv6 Internet is a truly viable 
> > > replacement for the community at large (enterprise, 
> > > education, end user, etc).
> > > 
> 
> [...]
> 
> > Sure, people who SERVE files out on websites and who make other
> > services available had jolly well best get their act together and
> > make sure their software runs on IPv6, pronto.  But I don't have
> > any sympathy for any of those people who are caught with their
> > figurative pants down.  They are being PAID by their customers to
> > know what they are doing!!
> 
> How the hell am I supposed to do that as a software developer when
> A) I can't get IPv6 space from ARIN and B) I can't get it from my ISP?
> 

Get a different ISP that can give it to you.

Unforunately, many ISP's are not being proactive and won't be until
they start losing customers.

As technical head of the ISP I work at, I can bring considerable pressure
to bear on our feeds to become IPv6 compliant.  Can I make it a do-or-die
proposition for them continuing to do business with it?  Frankly, I don't
know - I haven't tried.  But, suppose I can't - supposed if I tried I got
vetoed by the president of the firm.

All it would take for me to override the veto is for ONE of our
customers to quit our service, and tell us he was going to a competitor
who DID offer IPv6.   You see, customer loss is what the ISP president
pays attention to.

And it is only a matter of time before this happens.  It could even happen
tommorrow.  It will certainly happen to the ISP your buying service
from if you decide to tell them this.


> (I have a legacy class C, so I'm fine for IPv4.)
> 
> > 
> > Could you be more specific on this barrier of entry thing?
> 
> I've tried several times to get IPv6 set up on our LAN.  Each
> time I run into the immediate barrier that I can't get IPv6 address
> space.
> 
> Use the IPv6 equivalent of RFC1918, say some.  No, that's been
> abandoned, say others.  Get it from ARIN.  No, we're too small.
> (Under current rules, the most we could legitimately qualify for
> of IPv4 is a /24 (what we have), and since we don't qualify for
> a /22, we can't get IPv6 from ARIN.)  Sign the LRSA and start
> paying, and then you can get it.  No, not as far as I can tell...
> We are still too small.  Get it from your ISP.  Our ISP (Verizon)
> doesn't appear to offer IPv6 and there are 0 hits searching their
> web site for it.
>

Are you Verizon T1 or Verizon DSL?  If you're a T1 customer I
can tell you there's ISP's out there would would supply you with
IPv6 now.  If your on DSL then there are local ISPs that provision
over the Verizon DSL network who will likely be able to get you
IPv6 a lot sooner than Verizon.

For a software development testbed there is nothing stopping you
from simply making up IPv6 addresses on your inside network.  You
won't be able to use IPv6 to connect to anything on the Internet
but you will be able to find out if IPv6 interferes with your program's
use of the network.  And you can connect to test servers via IPv6
on your inside network that are using the same made-up scheme.

Renumbering an IPv6 network is a snap so when your provider gets it's
rear in gear you should be fine.
 
> The only viable strategies seem to be 1) to make up something out
> of whole cloth and hope that when IPv6 eventually becomes viable
> and we get interconnected, the whole thing doesn't crumble to
> oblivion because there are way too many "pirated" addresses out
> there

Even if your BGP to your upstream, they have to accept and not filter
your IPv6 advertisements for your pirated IP's to be a problem for
someone else.  And until you have real IPv6 connectivity to the rest of
the world it doesen't matter if you accidently use an IPv6 range that is
assigned somewhere else since you won't be connecting to the Internet with
it.

Ted

> or 2) to wait for everyone else to do something first and
> eventually play catchup.
> 
> [...]
> 
> > 
> > The job that takes the longest is the one that is never started.
> > 
> > Step 1 is getting IPv6 DEPLOYED at all of the top and mid-tier
> > networks BEFORE IPv4 runout occurs.
> > 
> > I do not believe that it is possible to hold the moral high ground
> > here and on one hand tell people that they must deploy IPv6 now
> > because it's the right thing to do, and on the other hand, work
> > at prolonging IPv4.
> > 
> > When the day comes that every ISP in business can call up their
> > peers and say "Turn on native IPv6 routing to us" then I think
> > we can turn our attention to trying to make a supply of IPv4
> > available to the little small guys who have decided to get out of
> > the rat race, stay on IPv4, and just allow their ISP businesses to
> > gracefully die away.
> > 
> > Ted
> > 
> 
> -- 
> John Santos
> Evans Griffiths & Hart, Inc.
> 781-861-0670 ext 539
> 
> 
>