nanog mailing list archives

Re: Stupid Ipv6 question...


From: Leo Bicknell <bicknell () ufp org>
Date: Fri, 19 Nov 2004 12:25:10 -0500

In a message written on Fri, Nov 19, 2004 at 05:15:26PM +0100, Lars Erik Gullerud wrote:
While that would seem logical for most engineers, used to /30 or /31 ptp
links in IPv4 (myself included), that does not in fact seem to be the
way things are currently done in IPv6, unless something changed (again)
while I wasn't paying attention...  /64 is the minimum subnet size, even
for ptp-links - there was even an RFC published relating to the use of
/127's (or, should I say, the recommendation to "don't to that"), namely
RFC3627 (aka "Use of /127 Prefix Length Between Routers Considered
Harmful"). But, you can still get 65536 ptp links out of a single /48 of
course.

FWIW, my test networks have always been configured with /126's, and
have never had an issue.

With the exception of auto-configuration, I have yet to see any
IPv6 gear that cares about prefix length.  Configuring a /1 to a
/128 seems to work just fine.  If anyone knows of gear imposing
narrower limits on what can be configured I'd be facinated to know
about them.

-- 
       Leo Bicknell - bicknell () ufp org - CCIE 3440
        PGP keys at http://www.ufp.org/~bicknell/
Read TMBG List - tmbg-list-request () tmbg org, www.tmbg.org

Attachment: _bin
Description:


Current thread: