Having said all of that, I did discuss the matter with Jim Gray before his untimely demise (this was before Madison). We discussed what was the canonically correct way (ignoring tedious concerns like performance, hardware costs, etc.) to structure a data warehouse. He was in no doubt that the relational core, with no OLAP data marts, was the best approach. In other words, the way in which Teradata stores and manipulates data today. But (to be sure I don't misrepresent a man I hugely admired) he DIDN'T say that we all should be doing it this way at present. In addition, he most certainly did not say that Microsoft's current approach was incorrect for the market it addresses. He pointed out that, in the current world, a core relational data warehouse (with no OLAP marts) is typically very expensive to implement, which is why many people use a dimensional one and why people use OLAP data marts. But, at some point in the future, when hardware (and specifically MPP) is cheaper, then that will change. And we can see Madison as, perhaps, the first steps along that road.
Dig Deeper on Database management system (DBMS) architecture, design and strategy
Related Q&A from Mark Whitehorn
Here's a guide to primary, super, foreign and candidate keys, what they're used for in relational database management systems and the differences ... Continue Reading
The unstructured data types common in big data systems are often better managed by a NoSQL database than relational software, Mark Whitehorn says. Continue Reading
Expert Mark Whitehorn explains what skills are required for predictive modeling -- and whether business users can do the work of data scientists. Continue Reading