Skip to content

Capacity Forecasting: What You Need to Know

We all know capacity management is one of the most critical process that lays one of the backbone for stable IT infrastructure operations addressing the capacity & performance issues.

In this article, I would be focusing on those grey areas of capacity management which is not covered in ITSM best practices and that is CAPACITY FORECASTING.

As per ITSM best practices, capacity forecasting is an activity which is done by a separate process called Demand management. Demand management uses PBA’s & UP’s (Patterns of business activity & User profiles) to understand the demands.

But in real IT operations, there are very few companies who have a demand management process and function to analyze the demand and estimate the future capacity needed. So finally, it comes to the technical people/ architects/ SME’s who has to play the role of Nostradamus to do the capacity predictions (with respect to storage capacity, servers, space, etc.). 

What is capacity forecasting?

Capacity forecasting is an activity used for predicting future demand based on past demand information with respect to space, labor, & IT equipment and computing resources.

Capacity Forecasting is very necessary to any IT service provider, since forecasts of future demand will determine the capacity (with respect to IT components, space, etc.) that should be purchased, produced by the OEM, and shipped to the service provider.  

Capacity forecasting can be classified into 2 types, Short term forecast & Long term forecast. Short term forecasting is very helpful for operational day to day activities. It projects the demand for next few weeks, month or months but less than a quarter. Long term forecasting is used for long term planning which generally projects the demand for half a year or more than one year. 

Why do we need capacity forecasting?

  • It helps us to do better long range planning, anticipating capacity requirements before business is impacted
  • It helps in diminishing the risks associated with new/ changing services & also reduces the system incidents and problems.
  • It helps us to adhere with the allotted budgets & control costs, spending money on the right things at the right time.
  • Since simple predictions & assumptions on IT capacity, generally goes wrong. 

Approach for capacity forecasting

Capacity forecasting can be done be done through either qualitative & quantitative methods.

Qualitative methods are based on subjective opinions from capacity & storage experts. This can be done using methods like:

Grass roots:  Here the capacity and storage demand is estimated by asking the capacity & storage administrator & experts.

Delphi method: Here the capacity and storage demand is estimated by asking the stakeholders in the capacity team & storage team but with concealed identities. 

Quantitative methods are based on mathematical & analytical techniques.

Causal relationship analysis: Here the Capacity and storage demand is estimated by statistical techniques to establish relationships between storage quantity and demand.

Trend analysis: Here the capacity and storage demand is estimated by checking trends with respect to seasons, months, etc. in a random way.

One of the most effective & accurate estimation is time series estimation, there can be two types of moving average models: simple moving average and weighted moving average. 

Simple moving average estimation

formula

  • t   is the current period
  • Ft+1    is the forecast for next period
  • n  is the forecasting horizon
  • A  is the actual demand from each period

 For example:

MonthCapacity storage in TB
Jan1,325
Feb1,353
Mar1,305
Apr1,275
May1,210
Jun1,195
Jul?

 

formula2

Weighted moving average estimation

formula3

  • t  is the current period
  • Ft+1  is the forecast for next period
  • n   is the forecasting horizon
  • A  is the actual sales figure from each period
  • w  is the importance (weight) we give to each period

For example:

MonthCapacity storage in TB
Jan1,325
Feb1,353
Mar1,305

formula4