BeyeBLOGS | BeyeBLOGS Home | Get Your Own Blog

« History In the Making | Main

April 14, 2009

Splash at Enterprise Data World

I was fortunate to attend the Enterprise Data World conference in Tampa, Florida. It always gets my energy levels up when I visit with a passionate group of people pioneering the data management world.

During the conference I presented a topic that is near and dear to my heart, Physical Architectures for Data Warehouse Workload Optimization. The presentation brings together concepts found is recent articles that I have had published; Degrees of MPP in Information Management and Building a Best-Fit Data Warehouse in TDWIs BI Journal(requires TDWI registration).

The heart of the presentation reminds data architects to clearly understand user behavior and workload before starting the data design process. Where we used to focus on data modeling techniques to match user workloads, e.g. Dimensional modeling is great for metrics based analytics; in todays world you also need to consider how the underlying physical infrastructure contributes to the user experience. This is not what we traditionally call physical data modeling where we adapt the models for the specifics of a given database vendor. But rather should this data model benefit from a shared-everything or shared-nothing database architecture. To download the presentation please visit http://www.dataupia.com/EDWpresentation.

It is always great when you present and have an interactive audience, great questions, discussions and a few a-ha! moments. Here are some of my favorites:
What is big data? Greater than single digit terabytes? Double digit terabytes? How about the point where the old rules for accessing data have to change.
The Cloud is simply a SMP server on demand for as long as you need it and not about scalability or GRID.
Major rule changes are needed as we transition from transactions to working with data sets and also when we go from small to big data.

I can hear them announcing the boarding of my flight home. As I sign off, I am struck with the simplicity but importance of my key take away from the conference; the old rules for data management do not apply when dealing with big data. How do you define Big Data?

Posted by BeyeBLOGS at April 14, 2009 12:30 PM

Comments

Post a comment




Remember Me?