Data warehousing as a science vs. an art form. The latter can be beautiful to look at, and even useful, but only the artist knows how to (re) create it.
Agile, TDD and continuous integration are mature concepts in software engineering, but for myriad reasons many data warehousing shops have continued to use the old ways. If it’s not broke, don’t fix it, right? But nearly every shop I’ve worked with is fighting one or more of the following issues;
- QA resource and scheduling issues, especially around regression and major release testing.
- Converting daily/weekly/monthly ETL frameworks to near real-time.
- Spinning your wheels in a dev environment with scant, broken or misleading test data.
- Database environment consistency, including both platform and application logic.
- Releasing unknown or un-ready code to production servers, despite a cripplingly slow and manual change-control process.
Data Warehouse Robot re-imagines the decision support system as a fluid, automated, evolving set of software objects that cache, conform and copy data into structures that evolve at the pace of business.
Interested in a discussion? Let’s get the ball rolling…