What solutions exist for the empty PRMD and O problem?
Michael.R.Storz@lrz.lrz-muenchen.d400.de Thu, 27 January 1994 18:28 UTC
Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa08700;
27 Jan 94 13:28 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa08696;
27 Jan 94 13:28 EST
Received: from mercury91.udev.cdc.com by CNRI.Reston.VA.US id aa13314;
27 Jan 94 13:28 EST
Received: by mercury.udev.cdc.com; Thu, 27 Jan 94 12:26:42 -0600
X-From: Michael.R.Storz@lrz.lrz-muenchen.d400.de Thu Jan 27 12:26 CST 1994
Received: from wally.udev.cdc.com by mercury.udev.cdc.com;
Thu, 27 Jan 94 12:26:26 -0600
Received: by wally.udev.cdc.com; Thu, 27 Jan 94 12:29:46 +0600
X400-Received: by /c=us/admd=attmail/prmd=cdc/ ; Relayed ;
27 Jan 94 12:25:12 -0600
X400-Received: by /c=US/admd=/prmd=XNREN/ ; Relayed ; 27 Jan 94 18:25:41 Z
X400-Received: by /c=de/admd=d400/prmd=dfnrelay/ ; Relayed ;
27 Jan 94 18:26:30 Z
X400-Received: by /c=de/admd=d400/prmd=lrz-muenchen/ ; Relayed ;
27 Jan 94 18:23:12 Z
X400-Received: by mta MTAwally in /c=us/admd=attmail/prmd=cdc/ ; Relayed ;
27 Jan 94 12:29:33 +0600
X400-Received: by mta cdc.us in /c=us/admd=attmail/prmd=cdc/ ; Relayed ;
27 Jan 94 12:25:12 -0600
X400-MTS-Identifier: /c=de/admd=d400/prmd=lrz-muenchen/ ;
940127191837605-MTALRZVEE
Content-Identifier: 940127191837605
Content-Return: Allowed
X400-Content-Type: P2-1984 ( 2 )
Conversion: Allowed
Original-Encoded-Information-Types: IA5-Text
Priority: normal
Disclose-Recipients: Prohibited
Alternate-Recipient: Prohibited
X400-Originator: Michael.R.Storz@lrz.lrz-muenchen.d400.de
X400-Recipients: non-disclosure: ;
Message-Id: <940127191837605-MTALRZVEE*/c=de/admd=d400/prmd=lrz-muenchen/o=/ou=LRZ/s=Storz/g=Michael/i=R/@MHS>
Date: 27 Jan 94 18:23:12 Z
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Michael.R.Storz@lrz.lrz-muenchen.d400.de
To: Receipt notification requested <mhs-ds@mercury.udev.cdc.com>
cc: Return requested
</c=DE/admd=D400/prmd=UNI-GIESSEN/o=/ou=HRZ/s=Fock/g=Falko/@wally-gw.udev.cdc.com>
Return-Receipt-To: Michael.R.Storz@lrz.lrz-muenchen.d400.de
Subject: What solutions exist for the empty PRMD and O problem?
More than a month has passed since Kevin Jordan made his proposal how the
problem with the empty organization and empty prmd could be solved. No real
discussion has been done in the meantime. Therefore I try to get the discussion
moving again.
To find a solution we have to recap how RFC 1327 maps addresses between X.400
and RFC 822, what it does special for the cases above and how this functionality
could be achieved with a MHS-DS based implementation.
RFC 1327 introduced
- a textual representation of X.400 addresses
- an ordering/hierarchy in the set of O/R address attributes
- a placeholder for missing attributes
- a format for lookup tables
This allowed mappings like
O$@.PRMD$@.ADMD$cdnnet.C$ca#cdnnet.ca# and
cdnnet.ca#O$@.PRMD$@.ADMD$cdnnet.C$ca#
The mapping itself was always a combination of a lookup and algorithmic part.
To map an O/R address with a omitted PRMD or O to a RHS encoding at least an
entry must be made in the tables which included the omitted attribute. This
empty attribute value is represented by '@'.
Now, how could this be resolved with Routing Trees?
I see the following possibilities:
1) Change RFC 1327 in the way omitted components are handled. Use the changed
algorithm for MHS-DS.
2) The RFC 1327 approach. I will call this the lookup approach.
This means all attributes must be represented in a O/R address routing tree.
If the value of a component is empty, use a placeholder. In the domain
routing tree the attributes associatedORAddress,
nonAuthorativeAssociatedORAddress and associatedX400Gateway ommitted
attributes must be specified, as the value a placeholder is used.
The mapping is done as described in RFC 1327.
3) The algorithmic approach.
Define how an algorithm can be specified, which will at least be able to
do the same mapping as defined in RFC 1327. But this algorithm may give you
more possibilities.
4) Use a combination of 2) and 3), that means use first a lookup and then use
the algorithm found with that lookup.
The algorithm found can specify how to map the whole address, or the rest of
the address not matched by the lookup.
Are there any other solutions?
My comments to these solutions:
1) I have no idea how this could be done. Even if there would be a way it would
mean to specify a new RFC and the gateways conforming to this RFC are
incompatible to the gateways of RFC 1327. I think we can forget this
possibility.
2) This would mean a change of the representation of O/R addresses in routing
trees. And a value for the placeholder must be found. The change of this
document would not be as big as a change to RFC 1327, because there are
only a few implentations out there. This solution would be possible to make.
3) This would require no change, but an extension to current definitions.
It must be defined how to lookup this algorithm.
4) This approach is the most powerful. It would resolve the problem and allow
more things to do. This is the approach suggested by Kevin Jordan.
Everybody should now comment, if there exists other solutions and which
solution seems to be the best to solve the problem. Then we can discuss how
the choosen solution could be defined.
--
Michael Storz
================================================================================
! X.400 : G=michael;S=storz;OU1=lrz;
Leibniz-Rechenzentrum ! P=lrz-muenchen;A=d400;C=de
Barer Str. 21 ! RFC822: michael.storz@lrz-muenchen.de
80333 Muenchen ! Fax : ++ 49 89 2809460
Germany ! Tel : ++ 49 89 2105 8720
================================================================================
- What solutions exist for the empty PRMD and O pro… Michael.R.Storz