Most of us have been charmed by circus jugglers who toss balls or bowling pins up in the air and keep them there. How do they do that, anyway? Juggling takes great concentration and good hand-eye coordination. Keeping objects in the air continuously requires planning and flawless execution. It seems to me that creating multiple OLAP cubes requires many of the same qualities as the act of juggling. Organizations in the forefront of achieving competitive advantage have already realized the benefits of generating on-line analytical processing (OLAP) databases, in the form of data marts, that address specific functional needs for multidimensional access to information. More and more OLAP cubes are being generated as we realize their capabilities to enhance corporate decision making. As we collectively address the need for multiple cubes, the issue of juggling those cubes draws attention.

Many cubes are required to satisfy different business constituencies and separate business problems. Some cubes invariably have similar dimensions to other cubes. Yet today we have to generate each cube and its set of dimensions separately. Unique generation means we have separate data transformation rules, along with separate extracts for each data mart required. Enter a new technology that incorporates the creation and population of multiple OLAP data cubes simultaneously. Juggling multiple OLAP cubes--creating and maintaining separate structures--requires planning and flawless execution.

Progressive OLAP vendors (who want to remain in existence beyond the business intelligence "maelstrom" that Gartner Group predicts will significantly reduce the number of players in the BI market) have or will soon announce capabilities for managing the creation of multiple OLAP cubes. One such vendor on the forefront of data mart management solutions is Decision*ism, with their product called Aclue. Pardon me for this aside, but doesn't this name sound like an advertising gimmick? I can just see an ad with a Valley Girl counseling us, "Don't be a loser! Get Aclue! " (We don't want to be "clueless," do we?) Regardless of the trendiness of the product name, however, Aclue provides the ability to directly populate and synchronize multiple multidimensional OLAP data marts in either Oracle Express or Arbor Essbase and, as such, represents a new wave in managing data mart implementation.

New products like Aclue allow the definition of many dimensions (say, 50 for example) as components. The user can then select a number of dimensions from the list of 50 that apply to the business situation at hand. The tool automatically generates the OLAP database structure, loading the dimensions and values without manual coding. While multidimensional databases in general have been criticized for their lack of flexibility, this new capability enables dynamic cube generation and the easy expansion of a multidimensional universe beyond what may have been originally envisioned. And, obviously, it decreases the workload necessary to implement new cubes.

However, lest you think these new tools constitute the constantly sought silver bullet, remember that there are downsides to everything. For example, cube creators spend much of their time not only designing the database, but sizing it to be as small as possible, thus allowing load, calculation and response times to be as fast as possible. Tools that automate cube creation don't necessarily address this issue (check with your vendor for sure); it remains a design issue to be dealt with. How long will it take to build the cube when an unknown combination of dimensions is selected? What will response-time performance be? And consider those 50 pre-built dimensions. While it's great to define them and make them available, it is nearly impossible to audit the thousands and thousands of possible combinations and roll ups for data integrity. Some can be audited, for sure. But to really take advantage of dimensional flexibility, you have to be very comfortable with your existing source data quality.

The data warehouse is still the foundation of any decision support environment for an enterprise. However, multidimensional OLAP cubes populated from the warehouse form the foundation for data marts that are usable by different functional areas of the enterprise to achieve rugged query performance. Juggling, or synchronizing, these cubes can become an implementation nightmare. And cube-juggling software provides an ability to manage the population of these cubes without manual intervention. What could be better? The whole idea behind data marts is to provide a view to data warehouse information that is consistent, yet unique to a particular functional area and that enables self-sufficiency among business professionals. These professionals will indeed be charmed by the juggling act that easily supports the inevitable frequent changes in their business. Concentration on data quality and practice in designing well-performing cubes using these new tools will provide that flawless execution we're all looking for in cube creation.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access