Here are a few interesting and unusual tools you can use to answer many critical questions in performance engineering, including some questions you probably found were never being asked at all.  Explore the tabs below to see some simple ones, and click on the menu at the bottom to launch a few tools that are slightly more sophisticated. (Note: To understand the inner working of these tools and to fully exploit them, you will need to license our methodology packs.)

The tabs below reflect a candidate sequence of usage in a typical IT project. First, you would use the Capacity Factor Calculator to size a deployment platform, and then to size the test platform as well. Next, you would use the Test Workload Planner to determine the performance test workload given the capacity decisions you have taken about the test platform. Having executed the test, you would check if the results are valid for further analysis by using the Test Validator. Finally, you would input results from a series of load tests into the Scalability Analyzer to determine if the application software is scaling well. (We have additional tools that plug in earlier in the lifecycle as well as later; these are available to methodology pack licensees.)

All tools featured here are Copyright © 2009-2015, Performance Engineering Associates. All rights reserved.

• Capacity Factor Calculator
• Test Workload Planner
• Test Validator
• Scalability Analyzer

The Capacity Factor of a system resource with a target configuration defines its processing capability with respect to a base configuration for the same system resource. For a given capacity factor, there may be more than one way to arrive at the actual deployment configuration. The two broad choices available based on the underlying support from the application architecture are scale-out and scale-up.

Projected increase in workload   % % Utilization Environment-> Base Observed value Base Threshold Target Threshold Capacity factor CPUMemoryDiskNetwork      Scale-upScale-out CPUMemoryDiskNetwork      Scale-upScale-out

*Practical ranges and limits of capacity factors need to be understood to use this tool effectively for the scale-up case.

While it is desirable to have the test environment exactly match the production environment configuration, this is seldom practical or possible. In most cases, the test environment is a scaled-down version and the workload targets have to be adjusted accordingly. This tool helps you determine the test environment target workload. The exact methodology and the steps are available in Methodology Pack WPT041.

Production Workload Target    arrivals / second Environment-> Test Benchmark Rating Production Benchmark Rating Test %Utilization Threshold Capacity Factor* Production to Test CPUMemoryDiskNetwork      CPUMemoryDiskNetwork      Computed  Test Workload Target     arrivals / second

*Practical ranges and limits of Capacity Factor needs to be understood for using this tool effectively.

Do you have a mechanism to objectively certify that the virtual users you specified in your load testing tool were actually created during the load / stress testing duration? This cannot be taken for granted on many occasions. This tool will do a quick check before you dive into the analysis of test results.

Virtual Users Specified Constant Think Time (seconds) Average Response Time (seconds) Average Throughput (requests/second) Test Outcome

Adding more hardware does not imply that your software can leverage the extra hardware and perform better. One of the key metrics that can quantify the way your software is scaling is the Scalability Metric, which this tool calculates. You need results from three independent load tests at different levels Lx.

Load Level Throughput (requests /unit time) Average Utilization (%) Computed Scalability Metric (w.r.t. base load) Base load L1 L2 L3

Note: Usually under high CPU and disk utilizations, the hardware contentions results in a disproportionate increase in system resource utilization that is not related to application workload performance or scalability problems. Similarly, it is important to keep in mind that under very low load conditions the comparative overheads due to resident software etc may become significant.

Disclaimer: The tools provided in the PERFORMANCE ENGINEERING ASSOCIATES (PEA) websites are provided “as is” without warranty of any kind. To the maximum extent permitted by applicable law, PEA further disclaim all warranties, including without limitation any implied or stated warranties of merchantability, fitness for a particular purpose, and non-infringement. The entire risk arising out of the use or performance of this product and documentation remains with recipient. To the maximum extent permitted by applicable law, in no event shall PEA be liable for any consequential, incidental, direct, indirect, special, punitive, recursive, or other damages whatsoever (including, without limitation, damages for loss of business profits, business interruption, loss of business information, personal injury, disruption of family life, or other pecuniary loss) arising out of this agreement or the use of or inability to use the product, even if PEA has been advised of the possibility of such damages.