[ppml] IPv6 flawed?

Stephen Sprunk stephen at sprunk.org
Fri Sep 14 13:08:46 EDT 2007

Thus spake "Leroy Ladyzhensky" <leroy at emailsorting.com>
> I know of some companies that may have 1000 plus workstations
> and servers on their internal network but would never qualify for
> IPv4 or IPv6 address's because that don't need much... 10 or so
> real IP's

Such a company would qualify for PIv4 and thus PIv6 because the policy 
counts _all_ hosts, not just publicly-visible hosts or NAT boxes.  You can 
get a PI assignment for a _completely disconnected_ network if you have 
enough hosts, though RFC 1918/4193 space is recommended in such a case.

> so, way down the road, when the move to IPv6 why would they
> number all their inside machines just because that's the block
> their ISP handed to them? just to renumber when they change to
> another ISP?

That's why we have PI policies for both v4 and v6.


Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking 

More information about the ARIN-PPML mailing list