[ppml] "Who's afraid of IPv4 address depletion? Apparently no one."

[Paul Vixie]

> >so there's an opportunity here for IPv4-rich ISP's to delay their
> >own IPv6 transition, including dual-stack, so as to acquire customers
> >and cash flows from IPv4-poor ISP's, during a transition period that
> >can be deliberately extended by such delay?
> 
> Interesting thought...  the answer is likely yes, but it actually may
> be more likely an opportunity for smaller ISP's as it's the ratio of
> supply over usage rate that determines how long one can stretch an
> IPv4-only connectivity model.  It changes to more of an opportunity
> for deeply-funded entities in the presence of a market-like transfer
> policies.

i wasn't saying large vs. small, which is often unrelated to ipv4-rich
vs. ipv4-small (think cogent).  in the absence of a market-like transfer
policy, ipv4-rich means latently-larger whereas ipv4-poor means latenyly-
smaller, unless and only unless the ipv4-poor start out so large that they
make a "voting bloc" that content providers will pander do by dual-stacking.

at the moment i don't see any ipv4-poor.  and, alain durand's mechanisms may
work so well that ipv6-only endpoints can work as well as ipv4-natted
endpoints work now, but with the added advantage of using ipv6 where the other
end speaks it (think gaming and peer-to-peer, where both-ends-natted is
painful, and where there is no content provider per se).

there may in other words be a soft landing in spite of the things we all
don't do, and the routing table growth will continue to be nonexplosively
incremental the way it has been for the last ten years.  the networks who
have to grow to live are not going to wait for IETF or ARIN to save them,
and they aren't going to come to ARIN with proposals that would save them,
they're just going to engineer their way around it, exactly as happened
with NAT.