A Characterization of the Sensitivity of Relational Query Optimization to Storage Access Cost Parameters
Frederick Reiss and Tapas Kanungo1
(Professor Joseph M. Hellerstein)
IBM and NDSEG Fellowship Program
Most relational query optimizers make use of information
about the costs of accessing tuples and data structures on
various storage devices. This information can at times be
off by several orders of magnitude due to human error in
configuration setup, sudden changes in load, or hardware
failure. We are interested the following questions:
-
Are inaccurate access cost estimates likely to cause a
typical query optimizer to choose a suboptimal query
plan?
-
If an optimizer chooses a suboptimal plan as a result of
inaccurate access cost estimates, how far from optimal
is this plan likely to be?
To answer these questions, we have developed a theoretical,
vector-based framework for analyzing the costs of query plans
under various storage parameter costs. We have used this
framework to experimentally characterize a commercial query
optimizer. We performed the characterization using database
statistics from a published run of the TPC-H benchmark and a
wide range of storage parameters.
1IBM Almaden Research Center
Send mail to the author : (phred@eecs.berkeley.edu)
Edit this abstract