Coverage metrics for actor model programs: A model-based approach

Context:

Software testing plays an important role in developing high-quality software. Developers need to understand software test effectiveness; therefore, using coverage criteria for test assessments is common. Selecting appropriate criteria for each software context is crucial, as it can lead to better fault detection. Each context requires a specific test suite to reveal faults, and developers should be aware of the context to select suitable criteria. In actor model programs, developers encounter a specific paradigm and need novel criteria.

Objective:

In this study, we utilize the Model-Based Testing approach and demonstrate its applications in actor model programs. Our goal is to identify coverage criteria using this method and evaluate its effectiveness.

Method:

We utilize a model-based testing approach to discover new coverage metrics in actor model programs. Initially, we identify the key properties of actor model programs. Based on these properties, we propose specific coverage criteria that can effectively expose faults. Lastly, we evaluate the effectiveness of these new coverage criteria through a series of experiments.

Results:

This paper shows that our suggested coverage criteria based on model-based testing can be more efficient than traditional coverage methods. Our experiments indicate that we can detect faults more effectively with these criteria. In particular, with the Pair of Event Type Receive Coverage, we can guide tests to find faults at least 20% faster than conventional tests.

Conclusion:

This paper uses the model-based testing approach for ensuring the quality of actor model programs. We define three new coverage metrics and evaluate them through experiments. These experiments demonstrate that the new coverage metrics can reveal faults more effectively than traditional coverage methods in certain situations.