Databases Reference
In-Depth Information
The Inmon/Kimball Religious Wars
The Kimball Perspective
What the Kimball advocates thought was at stake, in the mid-
dle to late 90s, was the difference between a cumbersome and a
nimble way of providing access to historical data. They thought
the issue was an either/or issue, a choice to be made between
data warehouses and data marts, with the correct choice being
obvious.
It is true that “nimbleness” was a major concern during those
years. Data warehouse projects were nearly always long-term,
big-budget projects. Like most such projects, they tended to fail
at a high rate. Most failures were probably due to the fact that, in
general, big complex projects produce big, complex products,
and that with increasing complexity comes increasing odds of
mistakes which, over time, often result in failure.
But some failures were also due to senior management losing
faith in data warehouse projects. As these large projects fell
increasingly behind schedule and rose increasingly over budget—
something large projects tend to do—the pressure increased
to produce results that had recognizable business value.
Patience wore thin, and many data warehouse projects that
might have been completed successfully were terminated
prematurely.
Against the background of failed data warehouse projects,
data mart projects promised results, and promised to deliver
them quickly. The typical difference in length of project was
about three-to-one: something like two to three years for the
typical data warehouse project, but only 8 to 12 months for the
typical data mart project.
And in fact, the success rate for data mart projects was signif-
icantly higher than the success rate for data warehouse projects.
For the most part, this was due to more modest objectives: one-
room schoolhouses vs. multi-story skyscrapers. It was part of
Kimball's brilliance to find one-room schoolhouses that were
worth building.
The Inmon Perspective
What the Inmon advocates thought was at stake was a “one
size fits all” approach vs. an approach that provided different sol-
utions for different requirements. Instead of Kimball's either/or,
they took a both/and stance, and advocated the use of opera-
tional data stores (ODSs), historical data warehouses and dimen-
sional data marts, with each one serving different needs.
