[27725] in North American Network Operators' Group
Re: Provider-local addresses?
daemon@ATHENA.MIT.EDU (Brian Candler)
Thu Mar 9 19:57:12 2000
Date: Fri, 10 Mar 2000 00:07:28 +0000
From: Brian Candler <B.Candler@pobox.com>
To: "Daniel L. Golding" <dan@netrail.net>
Cc: nanog@merit.edu
Message-ID: <20000310000728.B1037@linnet.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <Pine.BSI.4.05L.10003091435340.3857-100000@cartman.netrail.net>; from dan@netrail.net on Thu, Mar 09, 2000 at 02:49:03PM -0500
Errors-To: owner-nanog-outgoing@merit.edu
On Thu, Mar 09, 2000 at 02:49:03PM -0500, Daniel L. Golding wrote:
> I'm not sure I see the point. You can use the existing RFC1918 space for
> this kind of application, such as having fixed, non-changing addresses
> for DNS.
Except that you want to avoid clashes with your customers' own networks,
which may already be using RFC1918 behind NAT routers. You could hijack some
of this space ("if you connect to me, you may not use this particular subnet
of RFC1918") but that's not particularly friendly.
> The whole kettle of worms are downstream allocated IP customers - be they
> leased line, DSL, etc. This is the renumbering that can and does cost
> small providers many customers
Yeah, that's pretty inescapable :-)
Cheers,
Brian.
