[arin-ppml] Q1 - ARIN address transfer policy: why the triggerdate?

[Paul Vixie]

> Why do you feel it will be different from the commercialization of the
> Internet? In that era, the large telcos were the first to do something with
> IPv4 in the lab, but they were outrun by the smaller companies that got a
> product to market as quickly as they could. In the end, the large telcos did
> get their Internet services rolled out, and they mostly ended up buying the
> small fry.

this is a different history than the one i lived.  when uunet was founded, the
part of at&t research that had a connection to the internet, was not involved
in productizing the internet, and at&t had no commercial interest in the
internet.  at&t wasn't outrun, they weren't in the race at all in any form.

years later when i was operating CIX, i had to tell several large telcos who
joined up, about the difference between peering and transit, and how it was
that their CIX connection wasn't the only one they were going to need.  and
then there was the day i explained to QWEST why they couldn't use RIP, and
what BGP was.  (yes, this was the blind leading the blind, but there i was.)

> Why won't the IPv6 wave start with smaller more nimble companies who are not
> afraid of rolling out an IPv4-IPv6 gateway service that has a bugs?

that's a new topic.  for one, if that's how things progress, it won't matter
to the end result, since there will be no planned ipv6-only networks until the
larger networks are running dual-stack.  for two, the larger networks are the
ones who can predict their address space consumption accurately and will know
well in advance when the ipv4 pool will stop being able to support their
growth.  but let's stay on topic.

> > in this part of the analysis, there's good symmetry between "can afford to
> > be an early mover" and "is so large and with so much organizational
> > inertia that being a late mover would be fatal".  in other words the
> > smaller entities who can't afford to invest in ipv6 until there's an
> > in-year or in-quarter motive to do so, are also the ones who can do so the
> > quickest when their time comes, and the "dribs and drabs" of ipv4 space
> > that the big guys won't be able to use, will tide the little guys over
> > just fine.
> 
> Fair enough. But this implies that the small guys should not be waiting for
> the big guys (aka upstreams) to take the first steps. In fact, there may be
> a business case for IPv6 peering over GRE tunnels through an IPv4 upstream
> network as one element in a large bag of tricks.

there just is no in-year or in-quarter motivation for a small/medium sized
network to discard the last-mover advantage given the thin margins in this
industry.  adopting dual-stack depends on a "war chest".