[c-nsp] RFC 1918 on loopback?

[bill fumerola]

On Tue, Jan 15, 2008 at 06:07:41PM +0000, john heasley wrote:
> Tue, Jan 15, 2008 at 08:56:44AM -0800, Tony Tauber:
> > - Merger/acquisition/interconnection with another entity which uses them
> >   and there's an overlap.  ("That will never happen" are the words which
> ... which FUD is made of.
> 
> The dubious security argument and inter-AS debugging, such as traceroute,
> should be sufficient to end this discussion.
> 
> Need another?  BGP RID?

Need another? ICMP messages originating from these addresses and either
being filtered (by you, by others) or being ambiguous.

Need another? along similar lines of acqusition problems: networks of
your staff using the same space (home, coffee shop, offices, etc) and
VPN headaches (split horizon, overlaping routes) that result.

while there are many many reasons to move from RFC1918 numbering of
loopbacks and/or interfaces into assigned space, there are very little
reasons (mostly contrived or based on a false sense of security) to move
towards RFC1918 on devices.

RFC1918 defines them and everyone (read: Other People's Networks) treats
them based on those definitions.  except that treatment by OPNs is based
on interpretation and generally what suits the situation.

-- 
- bill fumerola / billf at FreeBSD.org