[arin-ppml] The non-deployment of IPv6

[michael.dillon at bt.com]

> > Multi-homing IPv6 will matter once IPv6-only networks get 
> rolled out 
> > and I don't see that happening for some time yet.
> 
> I disagree completely. The same mentality that is applied to 
> v4 resiliency should be applied all the same to v6. Not only 
> that, it is extremely easy to v6 multi-home, as there are 
> several very large ISPs who offer *free* peering/transit over 
> tunnels that you can use in conjunction with your existing 
> providers. [plug: he.net].

In addition, IPv6-only networks are already rolled out. The is 
one very well-known instance of an IPv6-only network run by Google
who will provide peering access to certain Google apps only on
IPv6.

Yes, we are in the early days of IPv6 rollout, but it is already
rolled out and carrying more traffic than the IPv4 Internet did 
in the late 1990s. Like any exponential growth curve, it looks
very flat in the early stages, but within a couple of years we
will reach an inflection point driven by IPv4 runout, and the
growth curve will gain a lot more visibility.

Note that it is not the actual IPv4 runout event that will drive
up IPv6 usage, but the fact that all major ISPs are targeting 
that time period for ramping up the stuff that they are currently
running in labs and in limited trials to gain experience with
IPv6.

People have been working at this for a couple of years now, 
knowing that sometime in 2011 they will have to support IPv6
for real. 

A nice exit strategy for a smaller ISP would be to develop
their support for IPv6 early, including all the operational
support systems and applications, databases, and training 
of 1st level support people. Chances are one or more of the
largest ISPs will run into problems with some part of their
IPv6 deployment, and the small ISP could sell the business
to the larger one simply on the basis of "been there, done that,
and we can help you get the T-shirt too".

--Michael Dillon