TAC Members Attending: Tom Barton, Mark Scheible, Keith Wessel, Jim Jokl, Janemarie Duh, Tom Mitchell, Steve Carmody, Eric Goodman
With: Tom Scavo, Kevin Morooney, Scott Cantor, Ann West
(AI) Mark will discuss the structure of the Attribute Release working group with Steering Chair Sean Reynolds. - DONE
(AI) TAC members asked to review the list of entities being filtered by MDA v7 to try to determine whether any of these are well-known or suspected highly used entities.
(AI) Tom Barton will check with Jim Basney on this, as well.
(AI) Ann or Kevin talk with Klaas Weirenga from GÉANT about a presentation to TAC meeting concerning their T&I roadmap for, say, the next 3 years? <= Deferred to August
Action Items from this meeting
(AI) Scott Cantor will post to the REFEDS list a summary of issues related to the slow signature verification on the Shibboleth SP (which may necessitate a faster move to per-entity metadata).
(AI) TAC members should review the guidelines for working group chairs and flywheels.
(AI) Janemarie will clean up the straw proposal for the 2018 nominations process.
See Ops Update 2017-06-22 for details about the deployment of InCommon MDA v7.1
Tom Barton requests that TAC (or someone) take up the issue of filtering policies with the eduGAIN Steering group to get agreement on policies that should be applied at that level, rather than having each federation choose which entities to filter or not.
Scott Cantor reported that signature verification on the Shibboleth SP is slow. If the aggregate doubles in size, it would take around 3 minutes to verify. This means the need for per-entity metadata may be more urgent. Perhaps more importantly, we need to have solution for discovery in place before (or soon after?) we begin the migration to per-entity metadata.
A new release of the Shibboleth SP software is scheduled for the end of this calendar year. If changes to the SP software are needed to accommodate a new approach to discovery, the requirements need to be documented now (or soon) to make it into the next release.
One (rather far-fetched) option is to port the SP codebase (or a portion of the SP codebase) to Java. The Shibboleth IdP software (which is all Java) exhibits good performance while processing large metadata files (and XML in general).
Charters have been finalized for the new working groups, but wiki pages need to be created. Some final tweaking may be necessary. Email announcements are needed. Can the WG sponsors coordinate the creation of the wiki pages and announcements? If the chairs are known, that can be communicated, as well.
Discovery - The current plan is to launch the discovery effort at TechEx. Scott Cantor recommends not waiting until TechEx to address the discovery issue since the window for making changes to the software is rapidly closing. (AI) Scott will post of summary of issues to the REFEDS list. There was discussion about whether adding this WG would tax the community capacity. The working groups currently organizing are non-overlapping, so it is thought that each will attract a different group of members, so capacity should not be a problem.
Call for Comments: There was also discussion about whether the Attribute Release WG would complement or conflict with the Deployment Profile WG. It was agreed to move ahead with a call for comments on all proposed WG charters and then evaluate the community response.
The guidelines for chairs and flywheels have been in place for awhile. (AI) TAC members should review the guidelines at their convenience. We should add the TAC co-chair approach to the document.
Janemarie sent a link to a Google doc including a straw proposal for existing TAC members seeking re-election. TAC discussed various pros and cons; for example, whether a current TAC member who self-nominates for another term can participate in discussions and vote on nominees. There will be times when members need to recuse themselves from certain aspects of the discussion. (AI) Janemarie will clean up the straw proposal.