Home >> Solutions >> IT Capacity Planning >> Pre-deployment Testing

Pre-deployment Testing

Applications must do more than simply fulfill functional requirements. Performance and scaling requirements are just as important and should be taken into account as early as possible during the product evaluation or development process. The goal is to be confident that new applications will meet IT service-level requirements and not fall short when launched.

The process for preparing critical applications for production should include steps for determining the optimal configuration for systems that will host the new applications, aligning application performance with business requirements, and taking data center architectural policies into account.

TeamQuest Software Addresses Predeployment Testing

Unlike many tools offering simple linear trending as a means for predicting future needs, TeamQuest Predictor incorporates analytical modeling. Analytical modeling provides a fast, accurate and cost-effective alternative to trending, brute-force testing, or discrete event simulation.

Analytical modeling extracts vital statistics from a baseline and builds a mathematical model of the system based on queuing theory. The baseline should contain a representative mix of scenarios and transactions, making sure to include those scenarios presenting the most significant load due to high concurrency or complexity. The baseline can be created by extracting statistics from an existing system or from a scaled-down test system. The test load can be created using a smaller group of users running the new application in a controlled manner or by using an automated load engine.

To provide for successful modeling, workloads are characterized, grouping resource utilization for business services into separate workloads during baseline creation. Adequately designed workloads, separating statistics pertaining to different applications or business services, will allow later evaluation of what-if scenarios based on the business activity forecasted for each workload.

Click for larger view

Once a baseline has been established, workloads have been defined and vital statistics have been extracted, analytical modeling can be used to simulate a vast number of different what-if scenarios. Among other things, the scenarios can include any of the following modifications to the system:

Transaction intensity by workload (application, service or business unit)
Number of users
Reconfiguration of hardware
Promotion or demotion of hardware resources

Click for larger view

Simulating the same diverse set of scenarios in a brute-force load-testing approach would call for extensive efforts in terms of software licenses, allocation of hardware resources and person-hours. Using the power of analytical modeling in TeamQuest Predictor, this becomes swift and inexpensive - one that won't limit the number of different scenarios you are able to evaluate.