2019-06-24 - Data Migration Subgroup Agenda and Notes

Notes from the meeting:

The meeting about migration at the F2F was unexpectedly large, and therefore focused on an overview of where we are, and an attempt to get necessary requirements to meet our needs. We need to settle on what type of implementation strategies will move us forward.

Ultimately, the Capacity planning team will develop a strategy for presenting a "Market Viable Product" by July of 2020. 

It was agreed at the migration meeting that adding the bulk load capability to existing APIs would be very helpful, though a tool kit is still necessary to do the heaving lifting - e.g. for mappings to connect different entities. 

Jon Miller remarked that "mapping legacy IDs to UUIDs won't be handled by the loader."

All systems have some kind of ID to link bibs to holdings to items. What is the staging to preserve those links?

There is already a lot of tooling that exists for data migration. This group needs to make an effort to gather tools and processes together to create a virtual implementers "binder." Perhaps some of the tools can be stitched together to make it easier. 

Texas A & M is undertaking a POC for a migration toolset based on the Camunda workflow engine. Jeremy Huff can demo this to the migration sub group some time in July. 

Requirements for APIs to handle bulk loading need to be written. Dale says there is a document that he will prepare for review at the July 1 meeting.

A discussion followed about whether to ignore the rather complicated bid/holdings/items data in favor of users, circ data, holds, fines/fees, organizations. But the difficultly is that things like loan rules etc. require a link to the item, which requires the data be migrated first.