Notion of session: picture

[George_Robert_Blakley_III@tivoli.com]

I got to thinking visually and produced the following .ppt diagram of what
the model for our specification
looks like in my mind.

I've included core tokens, how the flow from services to applications
through request/response XML protocols,
and how they are transmitted via protocol bindings from one application to
another.  I've also included in
dotted lines my notion of the two places where a notion of session would be
appropriate -- one in the
protocol binding layer (specified by us) and the other in the pre-existing
communcations protocol infrastructure
(specified by others and relied upon by us).

I've also distinguished visually between "tokens" which are just the raw
assertion, and "bound tokens" which
might also be signed etc... in order to protect them appropriately for
their environment of intended use (this was
the substance of my discussion about making signatures optional on the main
mailing list).

I hope the diagram is helpful.

I'm attaching both the .ppt source and a .pdf

(See attached file: OASIS-model.ppt)
(See attached file: OASIS-model.PDF)

--bob

Bob Blakley
Chief Scientist, Security
Tivoli Systems, Inc.

PowerPoint 4.0

Adobe Portable Document