Attending: Steve Carmody, Mark Scheible, Tom Mitchell, Janemarie Duh, Chris Misra, Keith Hazelton, Walter Hoehn, Scott Cantor, Tom Barton (22 min late!)
With: Dean Woodbeck, Kevin Morooney, Ian Young, Tom Scavo, Paul Caskey, Ann West, David Walker, Steve Zoppi, Mike LaHaye, IJ Kim
https://spaces.at.internet2.edu/display/inctac/Ops+Update+2016-04-14
Tom Scavo reported on an April 12 incident in which an entity descriptor for Leepfrog was un-published unintentionally. This cause an approximate three-hour outage, since it occurred at the end of the work day, until metadata could be restored. Investigation found that this was caused by an unexpected reaction by the web interface to a sequence of button push commands that caused the entity descriptor to be deleted. Discussing some ideas for fixing this.
Tom also reviewed the three incidents from two weeks ago; the two unresolved issues are related to the metadata signing process failing. Staff are confident they have a solution, but it is still undergoing testing and this is not yet in production.
The latest iteration of the charter is here:
https://docs.google.com/document/d/1wDgZXT-ia97QHoeN5b1LLQvHJ2qi06SrjVHx0WHMtTE/edit
There was discussion about the MDQ server, developed by Ian Young, that is a prototype. REFEDS has supported some of this work, but there is no organization supporting this now or for the future.
It should it be made clear in the charter that there is no expectation to take the current prototype to production. That may be one option, but the working group should explore others. If moving the prototype forward is a proposed option, InCommon should have a plan and estimated cost for development and maintenance of the software.
There is also concern that the scope of the charter is quite large.
(AI) TAC members should review (by April 28) the charter, particularly the list of 10 items, to determine if the scope is too broad, and/or if there are items that could removed from the list.
There have been two threads on the REFEDS email list this week related to attribute release and the issues over harmonizing R&S with EU privacy laws and policies. TAC members are encouraged to read that thread.
One issue emerged with UnitedID’s policy of requiring an SP to send a list of requested attributes, even if that SP is in R&S. UnitedID has now changed that policy. This led to a discussion about R&S and the concerns that trying to interpret R&S to comply with EU privacy requirements will lead to a dismantling of R&S.
Scott Cantor and Peter Schober have each developed proposals that revise R&S. TAC also discussed the need for an InCommon document separate from the R&S spec that provides advice and context for IdP site admins, so they can make reasonable and informed decisions.
(AI) Tom Scavo will check whether there is such documentation already in existence.
Original DRAFT set of thoughts
Comments
(AI) TAC members should review this list for completeness and appropriateness, keeping in mind that, while the items don’t necessarily need to be finished in 2016, they should be things that are important and on which we can make measurable progress in 2016.
Scott mentioned that a grant proposal involving the Shib Consortium (and submitted by Yubicode) has made the final round of the NSTIC proposal process.