Regression Test Suite Reduction for Cloud Systems
Oussama Jebbar, Mohamed Aymen Saied, Ferhat Khendek, Maria Toeroe · 2020
Cloud providers offer a wide variety of services to their tenants. Providers share large scale infrastructures to host their services and use configurable software customized with configurations to meet different tenant requirements. These configurations are often the main source of errors. Moreover, they undergo frequent changes, therefore, systems' compliance to requirements needs to be re-evaluated frequently using regression testing. The problem of regression test case selection has been extensively addressed in the literature, however, existing approaches do not tackle the problem from the configuration perspective. In this paper, we propose a configuration-based method for regression test suite reduction for cloud systems. Our method targets a set of faults summarized in a fault model, and it relies on a classification of configuration parameters based on their relation to the deployment environment. Our idea is that the relation of the configuration parameters to the environment can be explored to reduce the regression test suite.