A Capacity Planning Tool for Batch Parallel Processing Systems

A.F. Neto, J.R. de Oliveira, and R.S. Mendes (Brazil)


Simulation Tools and Languages, Performance Modelling, Capacity Planning, Discrete Event Simulation.


System capacity planning is a very useful method to predict which hardware will be necessary to run an IT system with high performance. The proposed tool in this paper is developed to ease this hard and important job, automating the analysis of the software in different ways. It uses the discrete-event type simulation based on a queue theory model to represent the hardware and software interaction. The model presented in this paper represents a simplification of any batch IT systems. In this type of system, commonly we have a lot of transactions that need to be processed by the system without human interaction and frequently use the benefits of parallel processing for better performance. We can apply these ideas to represent the workload in the simulation model as simple tasks that enter the system to be processed. Results show that it is possible to use this model to represent the real system with good fidelity.

Important Links:

Go Back