Data Warehouse Metrics
by Bill Inmon
Originally published December 16, 2004
Everything has metrics. Roads have speed limits. People have a weight. Cars have a tachometer. Days have temperatures.
Metrics help us organize our thoughts and make meaningful comparisons. Without even thinking about it, metrics allow us to size up an event or a condition. And data warehouses are no different in this regard. We need meaningful measurements of a data warehouse if we are to have a basis for comparing one company’s warehouse vs. another company’s warehouse. It simply is human nature to want to have a few meaningful measures that quickly tell their own story.
With that in mind, here is a list of metrics for the data warehouse environment:
Bytes for the data warehouse;
Index space; and
Number of ETL programs;
Frequency of execution;
Amount of data passing through the ETL environment; and
Where the ETL program is executed.
Number of data marts attached to the data warehouse;
Volume of data found in the data marts; and
Frequency of replenishment.
Data warehouse data:
Frequency of access per byte; and
Size of physical record.
Data warehouse structure:
Number of tables;
Number of rows per table; and
Average size of row per table.
By table, number of rows of data exiting the data warehouse;
By table, frequency of exiting; and
Criteria for egress.
Number of queries handled per day; and
Average amount of data of each query.
Length of time data resides in the data warehouse.
Other structure supported:
Adaptive data mart;
Near line storage; and
There you have it—a suggestion as to how to measure your data warehouse.
Of course this list can be expanded and altered in many ways. One problem with this list is that it refers to data only as of one moment in time. It is often useful to look at these metrics over time. So. you could add:
Or, you could section up the list by internal sub organizations. You might have something like this:
Often, it makes sense to logically group data together when the physical implementation does not have such a grouping.
Of course, you also could add the type of dbms the data warehouse is written in:
The modifications and additions are endless.
So, who is it that finds metrics useful?
SOURCE: Data Warehouse Metrics
Recent articles by Bill Inmon
Copyright 2004 — 2019. Powell Media, LLC. All rights reserved.
BeyeNETWORK™ is a trademark of Powell Media, LLC