10. BUSINESS ELEMENT FORECASTING › 10.1 Functional Description

10.1 Functional Description

The term "business element" refers to a unit of measurement
used to characterize activity in some division of the
organization.  The objective of Business Element Forecasting
is to investigate possible statistical relationships between
such measurements and the demand for computer resources.  If
you can establish such a relationship, you can build a
statistical model which relates the level of one or more
business elements to the amount of computer resources needed
to service that level of business.

Many types of organizations use the values of business
elements as a part of business planning.  For example, a
company which underwrites life insurance has an actuarial
department to study demographic factors such as age, health,
occupation, and geographic location to determine the expected
size of future insurance claims.  In addition, most
organizations attempt to forecast future sales volumes and
production costs as part of their budgetary planning.

Business Element Forecasting allows you to develop forecasts
that relate application resource consumption to business
elements that you have identified in your installation.
These forecasts are based on stepwise multiple regression
models.  You may apply the program to any of the resource
element history files that are produced by the capacity
planning database generator (see Chapter 3).

Once you establish a relationship between business elements
and capacity requirements, you can use the forecast data for
the business elements to drive the capacity planning process.
This allows you to base the data processing capacity plan on
the expected future workload of the organization as a whole,
rather than attempt to estimate future use only from the
perspective of historical use of the data processing
resource.

The program uses the SAS REG with the stepwise option to fit
a multiple regression model to the historical resource use
and historical business element values, perhaps suggested by
the users of the application.  The model is then used to
estimate the future resource requirements of the application,
using estimates of the business elements provided by the
application's users.  The SAS REG procedure is used to
determine the confidence limits for the forecasts.