Named Referrals Questions.

[Leonid Dubinsky <dub@shym.com>]

Dear LDAPEXTers!

Here are some questions/suggestions I have concerning
"draft-ietf-ldapext-namedref-00.txt".
Thank you in advance for your comments/clarifications.


QUESTION 1.

The draft seems to imply that objects with a 'ref' attribute can have
subordinate objects with a 'ref' attribute. This can be used to make
subordinate referrals more specific. Mark Smith confirms that this is
the idea. I think this should be more explicit in the text.


QUESTION 2.

The draft seems to also imply that objects with a 'ref' attribute can
have "normal" subordinate objects held by the same server. This can be
used to make superior referrals smart, i.e. to keep superior referrals
for each naming context held by the server in an object superior to that
naming context.

For example, if a server holds two naming contexts:

	ou=abc, o=xyz, c=us and
	dc=subd, dc=acme, dc=com

it can hold objects with the following dns that contain a 'ref'
attribute with the superior reference:

	o=xyz, c=us
	dc=com.

Such "smart" superior references can be used to provide referrals more
specific (and meaningful) than the default superior reference held in
the root DSE.

Mark Smith said: "I don't think that the idea of using referral entries
to provide more accurate superior references is discussed or implied by
the draft... perhaps it should be." He suggested to raise this idea on
the list. I do so now.

Mark also said that Harald Alverstrand sent a message to this list
recently in which he argued for more freedom for a server to decide what
referral URL to return, which is related to the "smart superior
references" idea. I failed to find this message in the archives.

I think that the text should elaborate on the intended use for the
referral entries that are superior to the non-referral ones - or
explicitly disallow them.


QUESTION 3.

The draft does not specify when (if ever) a server returns
NO_SUCH_OBJECT result to a client.

For example, if server A holds naming context "ou=abc,o=xyz,c=us" and
server B holds "o=xyz,c=us", it is reasonable for them to be glued
together by referrals - superior (smart or not) from server A to server
B and subordinate in the other direction. If server A gets a request for
the object with dn "cn=babs,ou=abc,o=xyz,c=us", the draft seems to
specify that referral to server B is to be returned.

I think that NO_SUCH_OBJECT result, and not a referral, should be
returned. Thin definitely needs to be clarified in the text - or am I
misinterpreting everything?


Regards,

Leonid Dubinsky			 Engineer
All opinions are mine, not my employer's.