Title: Intelligent
Metadata Model for Data Warehousing
Abstract: This paper
proposes a Metadata Model is designed to minimize utilization of resources
such as CPU, I/O, and Spool space when hundreds of jobs run simultaneously
as a batch processing in a four-hour time interval. The model lets the stored
procedures in different business and analytical subject areas run only when
source data has changed in the source subject area tables. Based on research
on one hundred business subject areas refreshes it was found that source
data change happened in forty percent tables. Skipping unnecessary loads
via this metadata driven approach not only helps save resources but allows
for a shorter cycle refresh time.