Lazy Product Discovery in Huge Configuration Spaces
This addresses the complexity and error-proneness in configuring highly-configurable software systems with thousands of interdependent options, offering a more efficient solution for software engineers and developers.
The paper tackles the problem of discovering valid product configurations in large, fragmented feature models with interdependent features, proposing a lazy product discovery method that shows significant performance benefits and succeeds where heuristic-based engines fail.
Highly-configurable software systems can have thousands of interdependent configuration options across different subsystems. In the resulting configuration space, discovering a valid product configuration for some selected options can be complex and error prone. The configuration space can be organized using a feature model, fragmented into smaller interdependent feature models reflecting the configuration options of each subsystem. We propose a method for lazy product discovery in large fragmented feature models with interdependent features. We formalize the method and prove its soundness and completeness. The evaluation explores an industrial-size configuration space. The results show that lazy product discovery has significant performance benefits compared to standard product discovery, which in contrast to our method requires all fragments to be composed to analyze the feature model. Furthermore, the method succeeds when more efficient, heuristics-based engines fail to find a valid configuration.