Information Technology Reference
In-Depth Information
Chapter 18. Capacity Planning
Plans are nothing, planning is everything.
—Philip Kotler
Capacityplanningmeansensuringthattherewillbeenoughresourceswhentheyareneeded.
Optimally this is done such that the system is neither under capacity nor over capacity. Re-
sources include CPUs, memory, storage, server instances, network bandwidth, switch ports,
consoleconnections,power,cooling,datacenterspace,andanyotherinfrastructurecompon-
ents that are required to run a service.
There are two major objectives of capacity planning. First, we want to prevent service
interruptions due to lack of capacity. Second, we want to preserve capital investment by
adding only the capacity required at any given time. Good capacity planning provides a
demonstrable return on investment (ROI) by showing the need for resources and capping
resource usage at a level that ensures good service.
Capacityplanningshouldbeadata-drivenprocess,usingdatacollectedabouttherunning
system to forecast trends. It also should be informed by future business plans for growth.
Thischapterexplainswhichdatayouneedtocollectandhowtouseittoforecastyourfuture
capacity requirements.
Capacity planning in large, fast-growing services becomes highly complex and relies on
sophisticatedmathematicalmodels.Organizationsoftenhireafull-timestatisticiantodevel-
op and maintain these models. A statistician with a technical background may be difficult
to find, but such a person is worth his or her weight in gold. This chapter introduces some
mathematical models that were developed for trading on the financial markets, and shows
how you can apply them to capacity planning in a rapidly changing environment.
This chapter does not look at the enterprise-style capacity planning required to meet the
day-to-day needs of the people who work for the service provider. Instead, this chapter fo-
cuses on capacity planning for the service itself.
Search WWH ::




Custom Search