Blog: Richard Hackathorn

« Review of Successful Business Intelligence by Cindi Howson | Main | Modeling The Future »

Data Warehouse Appliances - Where Art Thou?

As an analyst, it is fun to investigate a well-defined product area with vigorous competition. The positioning statements by vendors are often content-free and even humorous, reflecting many intense hours of debate. I understand; I have been there.

Data Warehouse Appliances (DWA) has been one of those well-defined areas. They are SQL boxes. Feed SQL statements in one end, and results stream out the other end. They are able to leap wide tables in a single scan, faster than a speedy join path, and so on. You get the point.

The marketplace has changed in the last few years. Those simple DWA products are not so simple, stretching our notions of SQL boxes. This category burring is a healthy reaction to market pressures.

In our DWA research study of 2007, Colin White and I recognized this trend toward higher and diverse functionality within the DWA marketplace. We create a new category called Data Management Appliance, which we defined as offloading data intensive operations from a host computer, such as operational, specialized analytics, or archival processing. Looking back, this was a bandaid on a much deeper issue.

That research did contributed the concept of an appliance as requiring:

* One Purpose – clear purpose
* One Package – tested, ordered, and delivered as a single system
* One Install – installed and maintained as a single system
* One Support – single point of service provided by a single vendor

This was amplified into the 9 dimensions of an appliance in a later article.

I am currently starting on the 2008 DWA Research Study, which will use this revised definition of an appliance. However, will the DWA label survive our scrutiny?

I doubt it at the conceptual level. DWA has historical value, popular recognition, and partial validity. However, the marketplace is definitely moving into an era of enterprise appliances that are evolving beyond SQL boxes.

The deep issue is how modular elements, like appliances (or whatever you wish to call them), should fit into the enterprise architecture. For several decades, enterprise systems were architected in an artistic fashion…a little piece here and another there. Some were truly works of art that even work sometimes. The world is changing too quickly to have that kind of artistic luxury. Besides the artists are getting old and retiring.

Data warehousing fits as a module (appliance) within the enterprise architecture. The challenge is to delineate the other modules (appliances) that will also fit into that enterprise architecture of the future.

Should I change the title of our research study to Appliances as Modules for Building Enterprise Systems? What do you think?

Share: