Capacity planning

What is capacity planning?

Capacity planning determines which resources an application needs for expected load and defined performance targets. These include computing power, memory, storage and networking. Microsoft describes usage data, demand forecasts and resource requirements as parts of this planning. For a new application, many values begin as assumptions.

Example: document jobs in a portal

In this fictional example, a team expects concurrent requests to generate PDFs. It defines acceptable waiting and processing times and plans trials with typical documents and particularly demanding ones. Expected jobs provide a planning basis; they are not evidence of load already handled.

Database access, file storage and the services on which processing depends also need checking. The number of processors alone does not tell you whether the time limit will be met. A load test should show where waiting times and errors arise and which resource actually limits performance.

Record assumptions, test conditions and measured results separately. If document size or complexity, usage or the performance target changes, the plan needs reviewing. Autoscaling may be one resulting measure; its limits and response time also need to be included.

Sources

• Microsoft — Architecture strategies for capacity planning