Thus wrote Mike Cheponis (mac@Wireless.Com):

> This appeared in the Sunday 27 Aug 00 edtion of the San Jose Mercury News.
> 
> of what is called network address translation (NAT). Since 1992, NAT has been
> deployed widely in the Internet. By any market measure, IPv6 completely lost
> out to NAT.

Yeah, that one again .. I work as netmaster in a sizeable European 
Local Internet Registry. I actually get to assign addresses, daily.
And I'm boound to RIPE rules which seem to be quite a lot stricter than
ARIN rules, from glancing at databases and current uses of addresses:

NAT has slowed down the need for addresses, but if you connect more
people, concurrently, you need more IP addresses for them. No more
addresses for printers and simple workstations in a LAN, and no more
addresses for shared webservers (i.e. forcing name based HTTP1.1) doesn't
mean we aren't going through address space like there's no tomorrow,
because so many -more- people flood on the net, and there's a limit on e.g.
how many shared webservers you can cram on a single system.

Also, I don't know of a single IP hostmaster who isn't sick & tired of
arguing with people about how they can't waste addresses and lay some on
the side for contingencies and pipe dreams either, but that they'll only
get what they actually need. 
With IPv6, you positively -wallow- in address space, someone says "here, 
this is my name and that's my organisations name, I need IP addresses" 
and you drop more addresses on them than they'll ever conceivably be 
able to use, no bickering about how one could use NAT and why NAT is not
possible in that situation and if they'll really go from ten servers to a
hundred in a year, just stick your head out and get more addresses than
all of IPv4 contains wound round your neck for that. :-}

Besides, IPv6 also incorporates lots of neat stuff as mandatory that's
a rare option in IPv4, and large installation admins will love it for that
as well.

regards,
	spz (currently proud user of 2001:680:1::/64)

PS: uhm, rant off :-)
-- 
spz@serpens.de (S.P.Zeidler)