Please note: This master’s research paper presentation will take place online.
Boren Zang, Master’s candidate
David R. Cheriton School of Computer Science
Supervisor: Professor Chengnian Sun
The adoption of model-based techniques for Android GUI testing has gained increasing recognition, emphasizing the use of formal models to define the application’s functionalities and GUI structure. This strategy enables the automated generation of test scenarios, ensuring a thorough exploration of the application’s various states and transitions. Ape is among the top performers in the domain of model-based Android testing tools.
However, Ape has limitations in testing efficiency and memory management. A notable limitation of Ape is that it conducts tests on a single device per testing cycle, suggesting potential enhancements in testing efficiency. In addition, Android OS has limited capability to handle memory management while Ape updates its model with increasingly complex states. As a result, the tool may experience slowdowns or fail to complete testing due to insufficient memory.
In this paper, DistApe is introduced as a distributed framework aimed at enhancing testing efficiency and addressing the issue of memory management through the coordination of multiple devices during testing phases. The assessment shows that DistApe reaches the saturation point 10%–25% faster than the original Ape, indicating its capability to improve testing efficiency.