Rapidly Changing Query Environments

Posted by: ward in Untagged  on

For small and medium size businesses, SMB, cubes are too burdensome. 

Mainly, there are too many skill sets required and the risk of being dependent on highly specialized people is too high a business risk and cost.   Even with the expertise, cubes are slow to design and build and the manufacturing overhead is large. 

For example, it can easily take a day just to install the OLAP software, then it can take days to design the dimensional model.  Then building the production processes to populate the cube on an ongoing basis can be a significant risk and expense.  Finally, once you have a cube being built, you need an MDX expert, and these people aren't very common and, after all these years that the industry has had to learn MDX, that says something about how difficult it is. 

Even with successful query results coming back from a cube, you have to figure how to render the results.  There aren't very many good OLAP clients that don't cost a lot, and the simple ones are probably not what you need for a dashboard or scorecard. 

If you're a solution provider or application developer deploying cubes for a living, then you will also have noticed that it's difficult to design and build cube related solutions remotely, so travel costs are higher than they need be, all detracting from the profitability of the solution.  

In short, I am suggesting that cube related costs typically overwhelm the cost-beneifit of most applications and many business needs are simply deferred or reduced in scope.

All in all, cubes involve quite a bit of effort and time.  If your original business questions change dramatically, it's all wasted. You may not need the cube anymore, but you may not able to drop the cube either, so your disposable IT budget is permanently eroded.

The net benefit of cubes are an almost factorial combinations of crosstabs.  But, in the same consideration, much if not most business information is now being delivered in a scorecard or dashboard.  Since a high degree of interactivity and user education is mutually exclusive with the goals of a dashboard or scorecard, I really don't see a benefit to having access to such a large number of crosstabs.  

In short, the benefit (of having lots of crosstabs) over cost of a cube (design, implementation and maintenance) isn't there.

 


Trackback(0)
Comments (0)Add Comment

Write comment
quote
bold
italicize
underline
strike
url
image
quote
quote
smile
wink
laugh
grin
angry
sad
shocked
cool
tongue
kiss
cry
smaller | bigger

busy