[69653] in North American Network Operators' Group
Re: Lazy network operators
daemon@ATHENA.MIT.EDU (Paul Vixie)
Sun Apr 18 01:45:15 2004
From: Paul Vixie <paul@vix.com>
To: nanog@merit.edu
In-Reply-To: Message from Paul Jakma <paul@clubi.ie>
of "Sun, 18 Apr 2004 00:10:06 +0100."
<Pine.LNX.4.58.0404180000050.22749@fogarty.jakma.org>
Date: Sun, 18 Apr 2004 05:44:38 +0000
Errors-To: owner-nanog-outgoing@merit.edu
> > ...
> > anyway, there will absolutely be NAT in ipv6 enterprise networks, but the
> > reason for it won't be a shortage of globally unique address space.
>
> Hmmm, or rather, there just wont be any demand for IPv6 deployment, at
> least from the edges (consumers, small/medium networks). Why bother
> changing if, despite the (almost indefinitely) availability of sparse
> address space, one can not claim a tiny piece as ones' own? Which is
> IPv4's only problem, at least as seen from the edges.
ipv6 solves problems that large providers have, or speculate that
they'll have, and at some point the edge/enterprise will adopt a dual
stack approach just to be reachable by 3G-etc, assuming that 3G-etc
ever stops having the 6to4/ISATAP/etc features it'll have to have
before it has v6 at all. so, theoretically, there's a tipping point
where dualstack or even v6-only will make sense even to someone who
refuses the lockins and stays with a NAT-based or proxy-based design.
> Provider independence (to some degree, even by DNS A6 or otherwise)
> for all should have been IPv6's biggest selling point. It doesnt have
> it, judging by multi6 it's not likely it ever will, and hence it's
> similarly unlikely there ever will be any real demand for v6.
provider independence is not a clear virtue in the eyes of those powerful
enough to sway vendors toward implementation -- that is, people with
billion dollar annual capital budgets. it may not even be a clear virtue
in the eyes of the edge/enterprise consumers who want to avoid lockin,
since they've gotten comfortable with NAT and proxies over the years, and
since in a post-CIDR world, everybody knows that the equilibrium between
can-route and can-qualify has to be more toward can-route and the people
who can't-qualify just need to make other arrangements.
i dearly wish i had understood this market equation at the time i was
pushing for A6/DNAME, since at that time i stupidly thought that this was
a technical problem and the people on the other side of the argument just
didn't understand the technical issues well enough. (oops.)
