Increasingly, we’re fielding queries about the EPR and how it behaves and how information about people (and about other business entities) will be integrated across the enterprise. This post is an initial attempt to capture some of the wider discussion.
The EPR is all (and only) about identity, and there is an infinite range of criteria by which any particular identity consumers might want to filter the identities they receive, and we’re not going to be carrying copies of every conceivable group in the EPR so we can make those routing decisions at the source. Instead, the EPR will simply publish messages about all people into the messaging infrastructure, from where the transform-and-filter routing plans will take over.
If it turns out, as a strange made-up example, that some downstream system only wants to receive a feed of people who:
- are older than forty years of age
- have received a B+ or better grade in ANTHRO 101 while studying between 2001 and 200
- attended Tikipunga High School
- were at any time between 2003 and 2009 employees of The University of Auckland in any organisational entity that is a child of Faculty of Arts
- have a registered Twitter account they have used in the last fifteen days
…then we’re only going to be able to know #1 from the EPR data — #2 and #3 will come from Student Administration, and #4 from HRMS, and #5 from outside the University.
Increasingly, those filtering decisions will need to be made by testing group memberships from within the routing plans inside the messaging infrastructure — that’s a normal feature of a publish-and-subscribe messaging, to separate fully the publishers from the subscribers.