2019-02-04 Meeting notes

Date

Feb 4, 2019

Attendees

  • @David Walker

  • Andrew Weiss

  • Curt Asher

  • Dana Ospina

  • Karen Schneider

  • Kevin Cloud

  • Renaldo Gjoshe

  • Tracey Elliott

  • Zach Vowell

Goals

  • Reach decision on single vs. separate campus repositories

Discussion items

Time

Item

Who

Notes

Time

Item

Who

Notes

5min

Recap

David

  • The need for multi-tenancy

    • Our other systems all use some kind of multi-tenancy

      • Alma, Primo, apps developed by the CO

    • DSpace

      • In the beginning: Separate instances, lots of problems

      • Now: Atmire-developed custom multi-tenant DSpace

    • Samvera Hyrax

      • Hyku was meant to fill this role

        • Overdue, still in-development

        • Upgrades after customizations would be difficult (compared to Hyrax)

      • Aaron’s model: Separate front-end instances, consolidated back-end

        • 25 prod servers + equal number development servers

        • Complex AWS, version control set-up

        • More flexibility than what we’ve previously had

        • Separate design, metadata, configurations

          • Some of this can be copied between instances, with attending overhead

  • Single instance alternative – see demo.

    • Pros:

      • Simpler set-up, can focus limited resources on this one instance

      • Uniform design, metadata, authentication

      • Improved discovery

        • But what about locally hosted campus repositories?

      • More easily accommodate systemwide initiatives?

        • Academic technology projects

        • AB 2192

    • Cons

      • Little flexibility for local metadata

      • Little to no flexibility for branding, institutional identity

        • Was listed as very important by campuses

      • Unknown scaling issues

        • likely addressed by AWS

      • Unknown issues around third-party integrations if not all campuses participate

        • e.g., DOI minting

      • Requires consensus on most decisions

        • Scope of the repository (see next issue)

          • Too narrow and we exclude content unnecessarily

          • Too broad and we swamp the search results with potentially irrelevant content

        • Metadata

  • Group is leaning toward single instance

    • But a more formal governance structure (before launch) and scope are critical

    • What voices are we not capturing in this discussion?

30min

Single vs. separate campus instance discussion (cont’d)

 

  • Group is leaning toward single instance

    • With caveats about the need for a more formal governance structure before launch

      • Scope of content is important in terms of hitting the ground running

      • Metadata is important here too

        • even more so in the digital archives

          • scope is broader

      • Should sketch out some ideas for governance

  • If Hyku were available, would we use it?

    • Are we leaning toward a single instance because of current technical / resource limitations?

    • Or are there good reasons to prefer a single instance over any type of multi-tenant option?

    • Can’t wait for Hyku

    •  

  • Are there good reasons to exclude any of the content currently in DSpace?

    • Institutional documents

    • Digitized historical materials?

    • Do we need a second instance for this other type of content?

20min

Secondary systems: Faculty profiles, open access publishing

David

  • Scope

    • Faculty profiles

      • VIVO (or other?)

      • Bespoke application?

    • Open access publishing

      • OJS now

      • Monographs?

      • Conferences?

    • Exhibits

      • Hyrax has a collections / exhibits collection

      • Spotlight another option

  • What are the costs associated with all these things?

    • We are under-resources centrally, so let’s not put too much on the CO plate

Action items

Decisions