Performance testing in continous integration environment

OData support
Supervisor:
Ujhelyi Zoltán
Department of Measurement and Information Systems

Throughout the development process of any software, it is essential for the developers to be aware of the current state of the product. Continuous integration provides the means for this: the code is built and tested automatically using a dedicated integration server and feedback is provided instantly for the developers. By following this practice, software faults and errors can be instantly detected.

However, it could prove to be equally important to constantly monitor the performance of the software being developed. For example, a sudden drop in performance is just as bad as functional errors. In such scenario it is crucial that all developers are notified immediately about the mistake, so the problem may be identified and corrected promptly.

This thesis studies the methods and tools available to realize performance benchmarking in Java environment and also examines the possibility of executing such benchmarks in a continuous integration environment to automatically provide feedback for the developers about the results. The developed technique is then demonstrated and evaluated through being applied to the continuous benchmarking of the EMF-IncQuery model query tool.

Downloads

Please sign in to download the files of this thesis.