*MACE-Dir-Metadirectories Conference Call*
February 27, 2002

*Participants*

Keith Hazelton -- Wisconsin(chair)
Tom Barton -- Memphis
Brendan Bellina -- Notre Dame
Steven Carmody -- Brown
Michael Gettes -- Georgetown
Mike Grady -- Illinois
Richard Jones -- Colorado
Bruce Krasnof -- Tufts
Steve Olshansky -- Internet2
Elena Ryazanova -- Tufts
Ellen Vaughan -- Internet2
Nate Klingenstein -- Internet2(scribe)

*Discussion*

        The MACE-Dir-Groups group has been adapted to begin work on
metadirectories.  This is expected to proceed for at least a month as
the group works to prepare Richard's draft metadirectory document for review
by the end of March.  The first role of the group was to analyze the
structure of the document and to work to accelerate the generation of content 
by enlisting help from other contributors.  Keith extended his offer to the
group to be facilitator for this process.

Lizard Brains

        The group first considered the way that the basic metadirectory
process could be partitioned into cohesive, useful sections.  Steven
summarized a metadirectory's basic role: data is fed in, a registry
aggregates it, and a mechanism then routes the data based on some simple
analysis to various directories or mechanisms using some simple
externally-generated provisioning rules.  The three significant
components in this process are the assembly of data, dealing with identity
management and the registry, and the big pile of consumers of the information 
at the end of the process.  The group was in favor of laying out this simple
model initially so that readers can understand how the pieces are intended to
eventually fit together as they parse the document.
        There are other aspects of that model which could be discussed
in the document as well.  The group mulled over how to include analysis of
how business practices and policies could be distilled into a set of
controlled rules on data propagation, as well as how to modify the business
practices themselves to make certain tasks vastly easier.  This business logic
could be represented in applications, the object registry, the ACL's of
intervening directory services, and many other places; the specific
implementation is typically based on where it is easiest to place, and
Michael called it nearly more of an art than a science.  Another
consideration is how to build the brains within the metadirectory,
handling identity conflicts and potentially contradictory data sources.
        A significant discussion about the implementation of metadirectory 
updating and controls on data ensued, where the group attempted to
determine how push and pull methods of obtaining information should be
handled by the metadirectory.  Tom explained that Memphis has a mixture
of both methods, picking one based on the situation.  When accounts change
state there, a grace period is allowed while e-mails are sent to the
account.  Other events leave information waiting to be picked up. 
However, this is not just a simple stimulus/response model; there is a brain 
in the middle which performs as a sort of nerve center where basic decisions
and analysis are performed.

Outline

        The group chose after some discussion to pursue a three-chapter
approach to the main body of the document.  There had been some suggestions 
of trimming the outline to have only two main sections, but the group
thought there was a natural overlay to discuss the three main components
as set out in the initial discussion.
        Keith also proposed that the group begin to work more extensively 
through the e-mail list to start to flesh out some of the finer points
in the document.  This initial development of the document will serve to
better place topics and to generate some initial text that can start
discussions and begin to form the document.