RE: Beginning taxonomy for finding LDAP servers.

[Ryan Moats <jayhawk@att.com>]

> At 14:21 1999-05-06 -0500, Ryan Moats wrote:
>
> >Method: Client configuration
> >
> >In this case, the client administrator configures it with a list of known
> >LDAP servers
> >to send queries to.  This list will be right (initially), but
> modification
> >to the list
> >requires client updates and doesn't scale real well.
>
> I agree, it doesn't scale. Therefore this only works if there are
> a limited number of known LDAP servers that a client has to be
> configured with in order to be able to find the rest
> (or at least the majority).
>
> Here in Norway, if everything works out as planed, we are going to
> put up a service which is going to contain basic information about
> every organization in Norway and if any of these organizations wants
> to publish more information about them self, through a publicly available
> LDAP server, a referral to that server will be stored in the central
> server/-s. Hence, potentially, you would only have to know about one
> or two LDAP servers, the second being the backup, i Norway in order
> to find every publicly available LDAP server in Norway.
>
> Worth noting is that we don't have to put any contraints on the
> DNs of the connected LDAP servers. Even though we are going to
> mandate the usage of one of:
>
> - the classical o=foo, c=no
> - the dc-naming dc=foo,dc=<TLD>
> - the guaranteed unique uniqueIdentifier=<organizationNumber>,c=NO
>

If I read this right, this is an implementation of the "referral method"
discussed in the mail between David, Jeff and I, correct?  If so, I think
it would be useful to discuss this as an example implementation in the
resulting I-D.  Roland, want to help author the thing (heh, heh :-)

Ryan