[82257] in North American Network Operators' Group
Re: The whole alternate-root ${STATE}horse
daemon@ATHENA.MIT.EDU (James R. Cutler)
Sun Jul 10 08:47:22 2005
Date: Sun, 10 Jul 2005 08:45:18 -0400
To: nanog@merit.edu
From: "James R. Cutler" <james.cutler@consultant.com>
Cc: "James R. Cutler" <james.cutler@consultant.com>
In-Reply-To: <20050709233546.A27909@cgi.jachomes.com>
Errors-To: owner-nanog@merit.edu
--=====================_37376274==.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed
Servers and zones are part of the physical instancing of DNS
roots. The definition of the root precedes the instance. In short,
the definition of a naming structure is disjoint from the delivery
and usage of that structure, be it servers and zones or whatever.
Regarding the separation of servers and zones, this is already common
practice. BIND provides a good example of this. With BIND, one can
serve an arbitrary set of zones from an arbitrary set of servers,
subject to the ability to do zone transfers. So there is obvious separation.
Cutler
At 11:35 PM 7/9/2005, Jay R. Ashworth wrote:
On Sat, Jul 09, 2005 at 06:08:25PM -0400, James R. Cutler wrote:
> Actually, many naming and addressing management experts consider that
> the existence of a root defines a unique namespace.
The existence of a root *zone* yes.
We really should separate root *servers* from *root* zones.
Cheers,
-- jra
--
Jay R. Ashworth jra@baylink.com
Designer Baylink RFC 2100
Ashworth & Associates The Things I Think '87 e24
St Petersburg FL USA http://baylink.pitas.com +1 727 647 1274
If you can read this... thank a system administrator. Or two. --me
-
James R. Cutler
james.cutler@consultant.com
--=====================_37376274==.ALT
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<html>
<body>
Servers and zones are part of the physical instancing of DNS roots.
The definition of the root precedes the instance. In short,
the definition of a naming structure is disjoint from the delivery and
usage of that structure, be it servers and zones or whatever.<br><br>
Regarding the separation of servers and zones, this is already common
practice. BIND provides a good example of this. With BIND, one can
serve an arbitrary set of zones from an arbitrary set of servers, subject
to the ability to do zone transfers. So there is obvious
separation.<br><br>
Cutler<br><br>
At 11:35 PM 7/9/2005, Jay R. Ashworth wrote:<br>
On Sat, Jul 09, 2005 at 06:08:25PM -0400, James R. Cutler wrote:<br>
> Actually, many naming and addressing
management experts consider that<br>
> the existence of a root defines a unique
namespace.<br><br>
The existence of a root *zone* yes.<br><br>
We really should separate root *servers* from *root* zones.<br><br>
Cheers,<br>
-- jra<br>
-- <br>
Jay R.
Ashworth &n=
bsp; =
&nbs=
p;
jra@baylink.com<br>
Designer &n=
bsp; =
Baylink &nb=
sp; &=
nbsp;
RFC 2100<br>
Ashworth & Associates The
Things I
Think  =
;
'87 e24<br>
St Petersburg FL USA
<a href=3D"http://baylink.pitas.com=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0" eud=
ora=3D"autourl">
http://baylink.pitas.com &nbs=
p;
</a> +1 727 647 1274<br><br>
If you can read this... thank a system
administrator. Or two. --me<br>
<x-sigsep><p></x-sigsep>
<font face=3D"Courier, Courier">-<br>
James R. Cutler<br>
james.cutler@consultant.com<br>
</font></body>
</html>
--=====================_37376274==.ALT--