[arin-discuss] neophyte IPv6 inital allocation and fee question

michael.dillon at bt.com michael.dillon at bt.com
Thu Jul 8 11:52:45 EDT 2010


> > Ideally you wouldn't totally ignore it but get your technical people
> to
> > do some experimenting with ULA addresses or a tunnel to he.net when
> they
> > have some spare moments.
> 
> i think a tunnel, to he.net or elsewhere, is an necessary immediate
> step
> even if it means renumbering when your physical upstreams gain ipv6
> capability.
> 
> i also think experimenting, either with ULA or tunnels, is a necessary
> immediate step for all providers, to be sure you've at least
identified
> all your non-ipv6-capable servers, routers, and business processes.

I disagree. I think that the activities which you describe, and which I
generally agree with, are necessary training activities for any
technical
staff who have no previous IPv6 experience. But I don't classify that as
"getting ready for IPv6" from the management point of view because it 
doesn't address products, marketing, transition costs or even detailed
project planning.

> > It is the national providers and large regional providers who have
> the
> > biggest risk in not being ready with IPv6 when IPv4 runout occurs.
> 
> while that risk does increase dramatically with footprint and capital
> plant,
> *noone* in this business can afford the risk of completely ignoring
> ipv6.

In any business there are always people who make good money for many
years dealing in obsolete stuff. I am certain that there will be many
ISPs who can survive and thrive with IPv4 only services, particularly
in larger markets where they have dozens of competitors.

> i'm reminded of the math for getting as many survivors as possible
from
> a
> burning building.  the most important ingredients are avoiding panic,
> and
> beginning the evacuation as early as possible, ideally before there's
> even
> an obvious need for urgency.
> 
> i strongly recommend that all of us get ipv6 working before there's
> urgency.

Evacuating a burning building does not consume resources that could be
better spent elsewhere. In fact, spending too much on IPv6 too early
could
be dangerous to the survival of a business. The art of handling the IPv6
transition is to spend the money just before you can earn revenue based 
on that spend.

National providers and large regional providers have a much longer term
planning horizon and habitually spend their money months before the
revenue
starts to flow. Those organizations have to act now.

But smaller businesses have a more delicate juggling act to deal with,
and
if their upstream providers haven't given them a firm data for IPv6
Internet
access, then they are better off waiting. They will reap the benefits of
the
work that larger ISPs are doing with vendors.

But at the same time, smaller businesses need to learn enough about IPv6
in
order to make the detailed spending and product plans, which is why that
they should get their technical folks to start deploying IPv6 today in a
trial phase, geared for learning and planning. This is the kind of thing
that
can be done as a low priority background process that only consumes a
few 
hours a week, can mostly be done without impacting business as usual,
and
therefore doesn't cost enough extra to bother counting it.

--Michael Dillon




More information about the ARIN-discuss mailing list