[arin-ppml] "Millions of Internet Addresses AreLying Idle"(slashdot)

[Paul Vixie]

> From: bmanning at vacation.karoshi.com
> 
> On Tue, Oct 21, 2008 at 07:33:45PM +0000, Paul Vixie wrote:
> > 
> > can you explain why someone would stop using IPv4 if it still reached
> > the entire set of endpoints they wanted to exchange traffic with?  or
> > failing that can you explain why someone would switch to IPv6-only if
> > it would limit the set of endpoints they could exchange traffic with?
> 
> 	thats a dirt simple question Paul.  Reduced complexity in the
> 	end system.  Check w/ anyone who has run large scale dual-stack
> 	infrastructure (like ESnet, DREN, SPAWAR, etc) and the one thing
> 	they pine for is less crap on the endsystem.  Running a single
> 	protocol stack is much simpler to install, troubleshoot, debug
> 	and audit.

did you miss the part where before this could happen there would be growth
of ipv6-only end nodes and thus ipv4 would be a nonuniversal protocol, NAT
or no NAT?  as before, i don't see why DREN would prefer non-RFC1918 IPv4
over RFC1918 IPv4 if neither one will have as many endpoints to talk to as
IPv6-only -- which is the point in the timeline we're discussing.