Just when you thought it was safe to catch your breath and hunker down with your general purpose DBMS data warehouse, data and business requirements are escalating everywhere past the point where this strategy is tenable. The market has filled this gap nicely over the past few years, providing us with a plethora of alternatives - none of which would seem to me to indicate a continued allegiance to a single platform enterprise data warehouse being the elegant sole end game. Database specialization is upon us and the primary determinant of an end client's portfolio is the workload.
End clients face an array of possibilities today to accomplish common objectives of getting leaner, more agile, becoming more real-time and reducing costs. The general purpose DBMS, hosted and supported in-house, will, of course, remain an anchor of operational and post-operational environments. Multidimensional OLAP will also continue to play a key role in many environments. Playing on the edges now in many environments are data warehouse appliances. Some have made it their enterprise data warehouse platform. There are memory resident DBMS. There are DBMS that store data in columns, instead of rows. Actually so many "analytic" specialist DBMS dot the platform today, such as NeoView, Aster Data, Exadata, ParAccel, and DB2 BCUs, garnering so much interest, that it would be unusual for a shop to not be considering one or more now.
All of these categories are not mutually exclusive! There are appliances that are columnar, memory resident and hosted in the cloud. Go figure. It's important to know the building blocks that solutions are made of, ensure they are all compatible with your workload and goals, and not look solely at how the platform delivers a single application.
Posted March 23, 2010 7:33 AM
Permalink | 1 Comment |