Takeaways from the World Wide Data Vault Conference
There were some really great sessions at the World Wide Data Vault Conference (#WWDVC), and I really need to take a bit of time to upgrade some of my templates to DV2.0. Thankfully that’s pretty straightforward with the framework in place already; the changes, or upgrade, can be automated as well using the same metadata.
My takeaways are;
- There is a push towards virtualisation for the Presentation Layer / Dimensional Model (or similar); using views directly off the Data Vault based Integration Layer model. This is in line with increasing performance / capability across the board
- It’s still too early to apply virtualisation on the Integration Layer level at this stage, due to performance considerations. One things we didn’t discuss but I’d like to pursue further is a different take on this; virtualising the Data Vault but capturing the Dimensional Model as a materialised / indexed view
- DV2.0 adjustments do not fundamentally change the concepts, but definitely make the ETL leaner and better performing
- There is a lot of emphasis on the symbiosis between standardisation and automation. You need standardisation to support automation, and automation verifies the standards and basically doesn’t work if the standards aren’t in place
- Lot of emphasis on automation, but in almost all cases based on automating SQL scripting
- The biggest hurdle in adoption ‘in the trenches’ is the staff training on-site, this seems to be a shared experience
Really good sessions!