Capacity Planning & Forecasting

Capacity planning is the process of adjusting the capacity.



Discrepancy between utilization and the demands results in an inefficiency, either in under-utilized resources or unfulfilled demand. The goal of capacity planning is to therefore minimize this discrepancy and balancing the demand vs utilization in an efficient manner .



The common view of the “Capacity Planning is not necessary“ as because most of the hardware is cheaper now a days. So people might thinking that they can afford the new hardware or upgrade there existing hardware as on required rather than going for Capacity Planning. Also they would like to go for excess capacity than the actual demand for the future increment of demand. It is very difficult to test the accuracy of performance predictions. So they would rather like to go for testing the performance to predict scalability.




In early sixties the hardware is too costly. Again up gradation is also costly. So the predication of capacity were mandatory those days due to cost factor.



As per Moore’s law processing power is tremendous increased. Powerful hardware is very cheap to affordable. Initially most of the system remains under utilized. So people thinking there is no need to plan for capacity.



The complexity of software increased day by day. In particular, current software (tools, frameworks, infrastructure, etc) is lagging somewhat in its ability to fully leverage the advances in hardware capabilities.



In today’s business scenario time is very much critical i.e. “Time is Revenue” . So downtime implies loss in revenue/business.



In today’s competitions the most common time horizons used in planning are one quarter. Management decision is considered as final for the business. And the business needs growth and early returns on investments. So the on demand capacity management must not inflate those predefined schedules. Also the manager only want a sense of direction rather than detailed “compass bearing”. There should be sufficient time for planning to procurements of hardware. Due to these factors nobody can afford the Capacity Management as on required basis.



A Capacity Planning model is not a model of any high tech engineered vehicle. It is like a road map. Just like a road map , we need to consider all those business matrices which are critical for capacity planning.



The question for Capacity Planning arise when there is large number of cluster node running in-house software applications and application performance is intolerable and erratic. In that case if it is necessary then the business might requires to spent millions of dollars in hardware.





The capacity planning question again arise when there is plan of doubling the workload due to increase in the business. The management might interest for the latent demand and time need to double the capacity.






The approach of Capacity Planning can be any one of these.Looking into the historical performance of the system from the past. Looking into the current utilization of the system in terms of %CPU or outgoing network data. Modeling the current data in regression analysis for a future prediction.



The techniques of Capacity Planning is basically divided into three categories. The first technique is the load testing while the system is in development phase. The second technique is the use of statistical analysis for trending. The third technique is the use of various modeling to predict the future.



Selecting the best techniques depends on the volume of data and accuracy needed.



There is two distinctions between the System management and Performance Management. While the system management consist of Backup/recovery, security and distribution of software where as performance management is monitoring the current, analysis of past and prediction of future.



The various layers of Capacity Management which includes past, present and future. We consider both the three phases in the capacity management , i.e Performance analysis of past data , monitoring of current data and prediction of future .




Capacity planning is the process of adjusting the capacity. Capacity Assessment is the process of proactively performance assessments. Capacity Forecasting is the process of predicting the future requirements.



The Capacity Management Tools must be sufficient enough to meet the following requirements.