Summary
Challenge: Conduct load testing for an accommodation booking service subject to significant peak demand each August, including repeated performance testing.
Solution: Used Micro Focus LoadRunner to simulate peak booking scenarios within a four-minute time window. Developed complex test scripts for the web-based application and executed baseline tests, followed by incremental load tests, before completing full-scale performance testing.
Outcome: Saved time by reusing scripts from previous years, delivered effective testing and recommendations, and increased confidence in system performance and reliability.
Challenge
AC were approached by a UK Top 10 University to carry out Load Testing on their Accommodation booking service. The application manages the allocation of student accommodation. It is subject to a major peak in demand every August, coinciding with the time when new students receive their A-level results and are offered places to study at the University.
An upgrade to the application has been performed since the most recent testing in 2018 (also performed by AC), and the University wished to ensure that the application can continue to perform acceptably. It therefore wishes to repeat the performance tests carried out in 2018.
The University also highlighted that the application only receives this level of traffic at predictable times of the year.
Solution
AC engaged with the University to further understand the specific requirements, focusing on the expected transactions based on previous years. It was determined that the University wanted to ensure that this application could support an estimated peak of 300 accommodation bookings within a 4 minute time period.
- AC had previously worked with the University on this application in 2018, utilising MicroFocus’ Loadrunner tool. Therefore, the decision was made to use LoadRunner order to take advantage of the work previously done in creating the scripts.
- The resource time saved by reutilising the scripts from previous years would offset the cost needed for the LoadRunner licences, as well as ensuring the tight deadlines were met. In future, AC may consider looking at OpenSource tools such as JMeter to complete this work without the need to purchase licences.
- The scripts needed to interact with the Web Based application. The scripts themselves required some complex logic, as each accommodation type only has a finite number of bookings that can be made against it. Therefore, when a type is fully booked, the script needs to select a new accommodation type.
AC completed baseline test runs as well as runs with a smaller transaction target to ensure confidence in the system as well as the functionality of the scripts.
Once satisfied, AC proceeded with the full load test, and repeated this test multiple times to ensure consistency in the results.
Benefits
Following the tests, the University could be certain that the application runs successfully, even during peak times.
Time-saving
The resource time saved by reutilizing the scripts from previous years would offset the cost needed for the LoadRunner licences, as well as ensuring the tight deadlines were met.
Successful testing
During initial runs, based on the results, AC made various recommendations to the University in relation to resources, and the application itself. These recommendations were implemented and tests were re-ran successfully.
Increased confidence in systems
The completion of these Load Tests ensure that the University can be confident that their system is capable of not only handling the expected load but also doing so in a manner that does not frustrate users with long loading times and failed bookings.
