[arin-discuss] ipv6 technology supplier phone bank?

Tony Hain alh-ietf at tndh.net
Mon Sep 28 15:58:43 EDT 2009

Paul Vixie wrote:
> > From: "Tony Hain" <alh-ietf at tndh.net>
> > Date: Mon, 28 Sep 2009 12:17:54 -0700
> > ...
> > > IPv6 is just the latest thing that ISPs are going to have to do
> that
> > > is going to cost more money, and deliver nothing in exchange.
> >
> > This is the short-sighted view of people that can't see past next
> > quarter.  Unfortunately most people are in this space, and will run
> right
> > up to the end of the free pool before they acknowledge that they need
> a
> > plan for the following quarter. People that insist on having an
> extremely
> > short term ROI will never understand the concept of avoiding a boxed
> > canyon.
> if one's competitors are doing short-horizon planning, then how can
> someone
> win new business without participating in this kind of "race to the
> bottom"?

Essentially this is the weeding process that leaves only those with deep
pockets standing. The IPv6 transition would have been relatively cheap for
everyone by spreading it out over time in a parallel deployment.
Unfortunately that didn't happen because short term ROI was the order of the
day, and will continue to be going forward. Once the source of IPv4 becomes
ebay, and the routing tables grow at an unconstrained rate, the costs will
be clear and the short term ROI will be to move to IPv6 as fast as possible.
For the ISPs this is a problem, because they will need to support IPv4 in
some form until they can motivate their customers to move. That means
raising the price of IPv4, but in the race-to-the-bottom mentality people
refuse to do that because those with deep pockets (allocation wise) are
again in the best position to hold out and drive competitors from the

At the end of the day, the CIOs and their architects need to earn their keep
by developing real 3-5 year plans that show the overall costs, and stop
ignoring the cost of keeping IPv4 running as 'the cost of doing business'.
Deploy CGNs (oh wait those cost money too, and the operational costs of
diagnostics will be much higher than people expect ...), break IPv4 as
people know it, then charge extra to put it back for those willing to pay.
In other words, make the price for IPv4 service align with the costs that
will be incurred to keep it running. Yes IPv6 deployment has a cost, but the
sooner it can be demonstrated that the cost for keeping IPv4 on life support
vastly exceeds that, the sooner work can begin on getting the replacement in


More information about the ARIN-discuss mailing list