Building the Data Warehouse

Скачать в pdf «Building the Data Warehouse»


D3—Performance Analysis


PRECEDING ACTIVITY: Data item set development.


FOLLOWING ACTIVITY: Physical database design.


TIME ESTIMATE: One week per subject area, unless the subject area is huge.


NORMALLY EXECUTED ONCE OR MULTIPLE TIMES: Once per subject area.


SPECIAL CONSIDERATIONS: This step does not need to be done in the case of small amounts of data and/or small amounts of processing.


This step is performed if the volume of data, the volume of processing, the traffic over the network, the growth of data and processing, or the peak period of processing will produce significant amounts of activity. If none of those factors will occur, this step is not done.


In this step, the issue of physical denormalization of data is addressed. Specifically, the design practices of merging tables, selective introduction of redundancy, creating popular pools of derived data, creating arrays of data, and further separation of data where there is a wide disparity in the probability of its access are considered.


If this activity is done at all, the output reflects a much more streamlined design, with little or no loss of the benefits of normalization of data.


PARAMETERS OF SUCCESS: A design that will be efficient to access and update, in terms of both data and programs that access and update data. Done properly, this step ensures efficient resource utilization.


D4—Physical Database Design


PRECEDING ACTIVITY: Performance analysis.


FOLLOWING ACTIVITY: Pseudocode development.


TIME ESTIMATE: One day per table to be designed.

Скачать в pdf «Building the Data Warehouse»