Baital: An Adaptive Weighted Sampling Platform for Configurable Systems

The diversity of software application scenarios has led the evolution towards highly configurable systems. Testing of such systems is challenging due to an immense number of configurations and is usually performed on a small sample set. Sampling is a promising approach for the sample set generation. t-wise coverage is often used to measure the quality of sample sets. Uniform sampling being most known method can fail to achieve high coverage in presence of complex constraints on configurations. Another challenge is a scalability hurdle for the t-wise coverage computation leaving sampling for higher values of t unexplored. In this work, we present Baital, a platform that combines two novel techniques for sampling of configurable systems. It is based on the adaptive weighted sampling approach to generate sample sets with high t-wise coverage. The approximation techniques for the t-wise coverage computation allow the consideration of higher values of t; they improve scalability for both t-wise coverage computation and sampling process