Re: using DNS for rfc1327 mapping rules, new I-D summary

Claudio Allocchio - +39 40 3758523 <ALLOCCHIO@elettra.trieste.it> Mon, 11 October 1993 18:16 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa24763; 11 Oct 93 14:16 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa24759; 11 Oct 93 14:16 EDT
Received: from mhs-relay.cs.wisc.edu by CNRI.Reston.VA.US id aa01659; 11 Oct 93 14:16 EDT
Received: from cs.wisc.edu by mhs-relay.cs.wisc.edu with SMTP (PP) id <28636-0@mhs-relay.cs.wisc.edu>; Mon, 11 Oct 1993 12:25:05 +0000
Received: from SYNW03.elettra.trieste.it by cs.wisc.edu; Mon, 11 Oct 93 12:24:52 -0500
Date: Mon, 11 Oct 1993 18:25:24 +0100 (WET)
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Claudio Allocchio - +39 40 3758523 <ALLOCCHIO@elettra.trieste.it>
To: huitema@mitsou.inria.fr, IETF-OSI-X400OPS@cs.wisc.edu
Cc: ALLOCCHIO@elettra.trieste.it
Message-Id: <931011182524.20800061@elettra.trieste.it>
Subject: Re: using DNS for rfc1327 mapping rules, new I-D summary

Hallo Christian,

>Original_From: Christian Huitema <Christian.Huitema@sophia.inria.fr
>Your draft is fine, but the use of the same syntax with two different meanings is
>quite confusing:
 
...
 
>I would suggest that you either use:
> 
>PRMD-inria.ADMD-red.X400.fr.   IN PTR inria.fr
> 
>or 
> 
>*.PRMD-inria.ADMD-red.X400.fr.      IN XA 10  inria.fr#PRMD-inria.ADMD-red.C-fr
> 
>In fact, I see why you would like to use the XA record -- having a "catch all" is
>nice.
>
>Christian Huitema
 
yes... using PTR has again the bad drawback of not being able to have a 
wildcard mechanism, and thus using XA also for "table 1" could be useful.

In fact we have to store both "table 2" and "table 1", which are not always
symmetric tables (I'd better not to make another statistic also on mapping
rules asymmetry... I could have other bad surprises, discovering how many
strane exceptions are around!).

I can see that your second suggestion can work:

*.PRMD-inria.ADMD-red.X400.fr.      IN XA 10  inria.fr#PRMD-inria.ADMD-red.C-fr

in fact the XA record value must just be a table 1 entry in reverse order
(822-domain#x400-or-name# instead of the canonical x400-or-name#822-domain# as
written in table 1)...

my doubt is: 

is it more difficolt to explain that the table 1 entry must be swapped when
insertedin DNS XA record...

or

is it more difficolt to explain that the XA value (on the right) is a table 1
entry, and thus it has by itself a reverse order?

Maybe it is more easy to use the second idea, stating something like:

"the syntax of the XA record is:

rfc822-domain-name   IN  XA  prio  rfc1327-table2-entry

and

x400-or-name  IN  XA  prio  rfc1327-table1-entry"

but personally I can also accept your original proposal, i.e. reverse
table1 entry when inserting the XA value.

above all, it is an info used mostly by s/w and not by humans.

thus, out there, which solution do you like more?

regards
Claudio