Forecasting, when modeling is not the only choice

When most organizations are tasked with forecasting changes within their IT environment, many immediately think they have to create models.

There are two types of modeling techniques that can be used when forecasting changes.

The first being Analytical modeling which basically looks at the workloads within an environment and measures the arrival of work into the server. This technique allows an organization to very quickly see the impact of business changes.

The other modeling, but time consuming, technique is Simulation modeling. Simulation modeling basically revolves around duplicating the existing environment and running synthetic or real world transactions at a certain speed to determine the impact of business changes.

Many times and for a number of areas, modeling does not fit into the forecasting of certain components. That is when using trending or stacked bar charts provides the largest value to understand the forecasting of those business changes. These are such as, but not limited to, consolidating servers, looking at the headroom available within a VMware cluster and disk space growth.

Key items many Capacity Managers want to know are:

• Do I need to forecast and manage this server or environment against peaks or averages?
• When are models appropriate for forecasting?
• When are trend and stacked bar charts appropriate forecasting?
• How to show both peaks and averages for a metric on the same chart

I discussed forecasting techniques which answered these questions and more at my webinar, why not register for our comuunity and listen to the recording? http://www.metron-athene.com/_downloads/podcasts/index.asp

Charles Johnson
Principal Consultant