Strategies for Analyzing Configurable Systems.
Alexander von Rhein, Sven Apel · 2015
Abstract: The advent of variability management and generator technology enables users to derive individual system variants from a given configurable system just based on a selection of desired configuration options. To cope with the possibly huge con-figuration space, researchers have been developing analysis techniques that follow different strategies to incorporate (static) variability. We discuss different strategies (variability-aware analysis and sampling) and evaluate them in different settings (model checking, type checking, and liveness analysis). A key finding is that variability-aware analysis outperforms most sampling approaches with respect to analysis time while being able to make definite statements about all variants of a configurable system. Generator-based approaches have proved successful for the implementation of config-urable software systems [CE00, AK09]. For example, the Linux kernel can be config-ured by means of about 10 000 compile-time configuration options, giving rise to possibly billions of variants that can be generated and compiled on demand. While advances in variability management and generator technology facilitate the development of config-urable software systems with myriads of variants, this high degree of variability is not