Attending: UAB - Jill, Jason;
UNC- Tyler, Nadim.
(1) Discussion of the two flow diagrams:
In
the one from CGU, there is no pass-thru so (if this is in fact how
the code is implemented) the commObjects must be publicly readable,
(including the passwords) which is insecure. Unfortunately no
one from CGU was in attendance; those of us on the call agreed that
the UAB flow on the left is correct, and that a separate conversation
should be scheduled with CGU on this topic. (Update - this'll
occur this week on Thursday).
(2) is a 'commOwnerDN'
attribute needed? The problem identified above was the reason for
bringing this topic up again. After discussion we agreed that
the answer was still NO. The pass-thru is handled by
configuration of the LDAP server; the application is aware only of
requesting and receiving commObject attributes. The LDAP
directory manager must set a rule (which can be based on use of some
locally define attribute such as commOwnerDN, or on matching the
returned DN with a substring of the commOwnerURI ...) In short,
this is an important topic for the cookbook.
Note to Jason -
make sure this topic goes onto the UAB to-do list for closing out the
cookbook.
(3) implementing endpoint authentication:
password-based implementations are difficult for endpoint developers;
they have to depend on directory servers being configured correctly,
etc. This seemed to be further compelling argument for some
certificate or S/MIME style approach.
(4) Discussion - do we
need an open source H.350 populating tooL?
Tyler suggested we do, to help with
deployment. To be a true software distribution (RPM) would be a
major development undertaking. In the meantime, UAB will
prepare some vanilla, re-usable PHP scripts that could be adapted by
those who know how to use PHP.
Note: these scripts are
available - Jason, can you let people know where to locate these
scripts?
(5) Room attribute? Is RFC 1274 for
real?
Cookbook needs to describe a scenario for needing a
standardized room attribute (note to jason - onto UAB to-do list).
The search field needs to be configurable so you can search by a room
name if you know it. We decided even if the RFC was old, it had
what was needed.
(4) other updates - Jill working on
presentations at NLANR (success) and CAMP (no success) workshops. ECS
beta testing is done. No further work on installing HCL proxy due to
problems no one could overcome & no reply from HCL.