Browse Prior Art Database

Complete a software performance measurement using cpu load rather than a prespecified number of simulated users

IP.com Disclosure Number: IPCOM000227829D
Publication Date: 2013-May-20
Document File: 2 page(s) / 43K

Publishing Venue

The IP.com Prior Art Database

Abstract

This solution proposes a new approach to completing a software performance measurement. This approach sets CPU load rather than a pre-specified number of simulated users to complete the measurement, therefore eliminating the trial and error process of getting to a specific CPU load during the measurement. Through monitoring the CPU load during ramp up of a performance measurement, this will determine the number of simulated users required to achieve the load requested by the user.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Complete a software performance measurement using cpu load rather than a prespecified number of simulated users

This disclosure outlines a new approach to completing a software performance measurement. This approach sets CPU load rather than a pre-specified number of simulated users to complete the measurement, therefore eliminating the trial and error process of getting to a specific CPU load during the measurement.

    To generate a specific CPU load using an automated performance testing product a user needs to add a specific number of simulated users. As the number of virtual users is decided before the performance run, getting a specific CPU load is a trial and error process which may be time consuming and wasteful of resources.

    Automated performance testing products examine test system behaviour and performance, by simulating hundreds or thousands of concurrent users to put a test application through the rigors of real-life user loads, while collecting information

from the test system. Performance tests will gradually increase the workload stress on the system until the product under test no longer behaves in a satisfactory manner. To achieve this, the performance engineer must increase the number of simulated users. Currently this needs to be done before the start of the test and the engineer needs to estimate the number of simulated users that will allow the target product to be placed under a certain stress for a predetermined period of time. The number of simulated users needed can vary depending on, code changes in the pre-release software, configuration and hardware changes. This results in time consuming system test environment set-up and significant performance test rework.

    It should be noted that a system's maximum acceptable capacity could be achieved before the hardware CPU load percentage reaches 100%. An experienced Performance Engineer will know how the product under test is likely to behave when the ha...