RE: draft-zeilenga-ldup-harmful-00.txt

I would say that the issue of whether multimaster replication is harmful or
not remains an open question.  In
http://www.ietf.org/internet-drafts/draft-slone-ldap-x500-align-00.txt
(summarizing the upcoming Proposed Draft Amendment (PDAM) on X.500/LDAP
alignment) section 3.11, I wrote:

   The PDAM will also call for comments exploring whether the existing 
   information model (X.501) requires modification in the presence of 
   multi-master replication (it appears that it does not require 
   modification) and whether the "otherStrategy" mechanism of the 
   existing replication protocol (DISP) is sufficient to signal the use 
   of external replication mechanisms such as LDAP changelog, LDIF, 
   DSMLv2, or proprietary mechanisms. 

In the PDAM itself (which is in the process of being registered for ISO
ballot), there are two editor's notes, which read:

Editor's Note 9-1: National Bodies and Liaison Organizations are requested
to provide comments on whether the existing DSE model can be used in the
presence of multi-master replication without modifying the DSE model itself.

Editor's Note 9-2: National Bodies and Liaison Organizations are requested
to provide comments on whether the use of the DISP "otherStrategy" mechanism
is appropriate for signaling the use of an external replication method such
as DSML v2, LDAP changelog or LDIF.

As best I can determine so far, the biggest impact to X.500 will be in the
procedures associated with distributed operations and chaining. At present,
a DSE can be of type "master" or "shadow".  Note that the DSE model doesn't
make a distinction between "a master" and "the master" -- simply "master".
The accompanying procedures (X.518) check to determine is the DSE is of type
"master" before processing modifications. 

What would be helpful to me is for another set of eyes to look at the
procedures, the model, etc., in an attempt to determine the answers to the
questions raised in the editor's notes. 

 -- Skip


-----Original Message-----
From: Kurt D. Zeilenga [mailto:[email protected]] 
Sent: Tuesday, May 06, 2003 11:04 AM
To: John McGarvey
Cc: [email protected]
Subject: Re: draft-zeilenga-ldup-harmful-00.txt



At 07:32 AM 5/6/2003, John McGarvey wrote:
>IMHO, this draft takes a very extreme position, stated here:

This draft only reiterates the position which RFC 2251 and
X.501(93) took, and which LDAPBIS is presently taking.  LDUP takes an
extreme position that an extension to LDAP can alter the LDAP service model
in ways which effect implementations not supporting the extension.  Such
proposals should not be treated as extensions, but as NEW protocols.

Kurt