– Scott Adams, director of product management at TeamQuest (www.teamquest.com), says:
Many organizations are dependent on servers to deliver crucial services in support of key business processes. For these organizations, when servers aren’t performing well, business suffers.
Despite the importance of managing server capacity, many managers operate in a reactive mode. They monitor for problems. When one is detected, there is a mad scramble to identify the cause and take corrective action. Oftentimes the extent of the proactive measures consists of guessing future capacity requirements and then padding those estimates as an insurance policy against future bottlenecks. This strategy is grossly inefficient, typically resulting in servers that are only 5%-10% utilized.
Whether you are talking about servers or other IT infrastructure, good capacity management is a journey. It takes IT managers from a reactive, firefighting atmosphere to a more mature environment where capacity planning prevents most performance issues from ever occurring. The time saved from eliminated unplanned work can be invested in new money-making applications and services.
Frequently, efficiencies can be gained by well-planned implementation of virtualization to consolidate servers. It can be possible to decommission large numbers of servers, for example, increasing the utilization of remaining servers while continuing to meet service levels as the business grows.
With the right tools, managers can be sure that the configuration changes they make will satisfy business goals without needlessly adding capacity that is overkill. And they can be sure that they aren’t adding CPU when what is really needed is a change in storage configuration or a minor change in software configuration.
An option that is being considered more and more by IT organizations is to allow peak workloads to be accommodated by tapping into cloud computing resources. This can be a way to avoid sizing internal systems to accommodate a level of growth that is most likely not to occur. The downside is that cloud computing introduces a whole new set of risks and financial variables to the capacity management equation.
The key to maximizing productivity, efficiency, and avoiding embarrassing outages is advance planning. Don’t assume that reactive processes and new dynamically scalable technologies will eliminate the need for careful capacity management.