Optimizing Software Development: A Systematic Approach to Software Performance Testing
Kwong, Kendrick (2024)
2024
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2024051713054
https://urn.fi/URN:NBN:fi:amk-2024051713054
Tiivistelmä
This final year project delved into the realm of software performance testing, focusing specifically on evaluating the quality of a service by examining response time and availability. A variety of metrics was employed such as mean, median, quartiles, variance, and standard deviation to assess performance indicators. The testing approaches encompassed load testing, stress testing and soak testing using tools such as Apache JMeter, InfluxDB MongoDB, Grafana, and the Robot Framework in conjunction with hardware and software components such as the RTLS Engine, Controller and Simulator.
The study presents reflections on the performance of the systems under examination. The findings of load tests, stress tests, and soak tests are discussed. The results of the tests are satisfactory. For the load test, it is discovered that there is a performance degradation when the simulated tags exceed 15k. For the stress test, it was found that the positioning data of the tags becomes unreliable when the simulated tags exceed 20k. The system was stable throughout the soak test although the response time of the request increased around 10% compared to the load test.
In conclusion, the results indicate that the performance of the application is satisfactory. Although bottleneck exists when the simulated tags increase beyond 15k, the whole system did not experience a performance creep. It is recommended to re-run in the future, ensuring that the application maintains its quality standards, particularly when integrating new software versions.
The study presents reflections on the performance of the systems under examination. The findings of load tests, stress tests, and soak tests are discussed. The results of the tests are satisfactory. For the load test, it is discovered that there is a performance degradation when the simulated tags exceed 15k. For the stress test, it was found that the positioning data of the tags becomes unreliable when the simulated tags exceed 20k. The system was stable throughout the soak test although the response time of the request increased around 10% compared to the load test.
In conclusion, the results indicate that the performance of the application is satisfactory. Although bottleneck exists when the simulated tags increase beyond 15k, the whole system did not experience a performance creep. It is recommended to re-run in the future, ensuring that the application maintains its quality standards, particularly when integrating new software versions.