Hey all,

For those who have attended the Collibra “Path to integration innovation” session yesterday, what are your impressions?

It was a really nice overview of all the integration capabilities, through edge, spring boot, apis, etc.
But I was a little disappointed regarding new capabilities and vision going forward, with some substantial pain points remain unaddressed.

Configuration in the UI

There was some teasers about "Making configuration easier through the collibra UI", but little details were shared on it. I think it mostly related to the edge configuration possibilities.
I see very compelling implementations on other technologies such as Sinequa, that allow the development of custom connectors in C#, directly available in the UI of the product. Having native capabilities with custom connectors, like salesforce, onetruth, etc. would be a blessing.

Metadata enrichment

There are many disconnected integration capabilities that make metadata enrichment scenarios impossible outside of custom development. You cannot combine manual edits with collibra lineage for example, since the integration will remove all manual edits. I'm looking forward to more details on what metadata will be retrieved from Collibra DQ and pushed to Collibra DIC, especially when there are data transformations. Stitching is still a difficult point as well.
To work around lineage limitations, we're looking at technical lineage platforms such as Octopai, but haven't figured out how we'll manage to combine metadata from multiple sources (edge, octopai, collibra DQ) unless we custom code everything.

Spring boot

I still don't understand what is the plan forward for the marketplace spring boot applications and how it will fit into Collibra's strategy. Usage of those application is for now relatively cumbersome, and we're thinking about our reference integration architecture. At this point, the applications are more often "templates" to show how we could develop custom integrations, but not production ready to be well integrated with the DGC platform.

Non-relational data

I am confused regarding how collibra should be considered. Sometimes it is presented as a "business perspective" on data targeted for non-technical users, sometimes it is presented as an end-to-end catalog that can serve different personas (architect, data engineer, etc.), but not clear how the different personas are served by the model and the capabilities.
In the call, Roberto mentioned the interest to catalog XSD files, that represent an XML data model. Those are typically non-relational by nature, so trying to use the jdbc connector for XML evidently does not meet expectations.
The same applies for many other topics: APIs, Kafka, json, mongoDB, Hbase, etc. I think it is somewhat misleading to think semi-structured data is supported when all those data models are reduced to a relational form. What is the vision of the future then? Will there be support for non-relational data (APIs and such)? Or will Collibra focus on a "business friendly", relational view of the data, leveraging logical data models?

Accelerating standard connectors

Many collibra customers reimplement the same integrations (Denodo, QlikSense, Informatica, etc.), which seems a relative waste of time compared to the possibility of having ootb connectors. I hoped the webinar would provide some visibility into a strategy to accelerate those integrations, but it seems the road ahead is still complicated.

So, what are your thoughts? Anything you really like or concerns?

Tagging a few people to get the ball rolling: @roberto.donato.bnc.ca, @david.botzenhart.duke-energy.com, @chris.allen.angloamerican.com, @evert.van.loo, @mahtab.rahman, @mudassar.abdul.teck.com, @stan.thompson, @martin.gall.huk-coburg.de, @peggy.mccoy, @teresa_gustavson.americancentury.com