In addition,
TPC-D focuses on
an important
class of database queries
called
Decision Support
(DS) queries.
Decision Support Systems
are designed
to help
management
monitor
key factors
in their business operations
by accessing
business databases.
DS Systems
might be used
in tracking
revenue,
sales trends,
inventory levels and
distribution problems,
and for doing cost and profitability analyses.
For example,
DS systems
are used extensively
and successfully
by Wal-Mart and
Dayton Hudson Corporation's Target stores
for inventory management
and other functions [Hershey 96].
The rest of the thesis is organized as follows:
Section 2
motivates and explains
query optimization.
Section 3,
introduces the
terminology necessary
to describe
the Cascades architecture.
People conversant
in optimization terminology
will want
to skim over
these sections,
just enough to see
the particular terminology
used in the rest of the thesis.
Section 4
covers
some
related work
on database query optimization.
Section 5
describes
the optimizer framework and
the model structure in Cascades.
Section 6
describes
Model D (D for TPC-D),
including the
selection of
logical and physical operators,
logical properties and rules.
Section 7
discusses the performance
of the Model D optimizer
on the TPC-D queries,
with respect to
extensibility,
efficiency,
quality,
and accuracy,
and how
rule selection
and
rule tuning
affect
the efficiency/quality
trade off.
Finally, sections 8 and 9
discuss future work and conclusions,
respectively.
Page 3