Capacity Planning Techniques

A variety of capacity planning techniques are available, each useful in its own right. TeamQuest provides tools for each of these techniques.

Simple trend analysis looks at historical data and projects a linear trend line into the future.

Can be accomplished without sophisticated modeling tools.

Does not consider non-linear behavior buried within most complex multi-tiered systems. Does not allow for experimentation with different configurations.

Can graph a trend line and can generate alarm events when performance varies significantly from a projected trend.

Stress testing typically involves sending synthetic transactions coming from imaginary end users to test systems to be certain they can support required service levels.

Can spot non-linear bottlenecks that would be missed using simple trending.

Stress testing can be time consuming and expensive. It can also miss important problems if the test environment is missing some workloads that would be present in the production environment or if the test systems are smaller than production systems. Experimenting with alternative configurations requires real changes to the test environment.

Tracks performance during stress testing so that you can tell if service levels are being met. TeamQuest software can also extend the results of stress testing, allowing you to predict performance using different workloads and configurations than might be available in the test environment.

Discrete event simulation, a more sophisticated capacity planning technique, actually simulates the queuing events that occur during execution.

Much more accurate than simple trending. Accounts for non-linear behavior. Allows for predicting performance on hypothetical configurations.

Can be very time-consuming and tedious to build and run the models.

Includes simulation modeling capability, though our customers generally prefer and TeamQuest usually recommends the use of our analytic modeling capability for most situations.

Analytic modeling uses mathematics to calculate how a queuing network will perform. To use analytic modeling, a baseline set of measurements is taken on a system and a model is then built and calibrated based on a description of the system. Once the model is built, hypothetical changes can be made to system configuration or business workloads, and the model will predict how the changes will affect performance.

Many of the same advantages as simulation modeling, but faster and easier.

One of the reasons that analytic models are simpler, faster, and easier to set up and use is that they are not completely generalized. If they were, they would suffer from the same drawbacks as simulation models.

Features a versatile analytic modeling tool called TeamQuest Model. TeamQuest Model can predict the performance of complex, multi-tiered computing environments that include systems from different vendors running a wide variety of operating systems. TeamQuest Model understands and takes into consideration a long list of variables including I/O configurations and CPU types and numbers. It understands multi-threaded processing and a growing list of partitioned or virtualized environments. Chances are, TeamQuest is the right tool for your capacity planning needs.