Metrics for optimizing post-release configuration of in-vehicle software

Abstract

Complex software architectures and designs to support more and more functionalities, higher software quality and faster time to market requires an optimal decision making. A measurement framework, which is developed based on Goal- Question-Metric (GQM) approach is proposed in this thesis to be applied in decision makings about software download. The framework is designed through a case study at Volvo Cars Corporation. The goals for decision making are quality of software product and efficiency in processes. The relative attributes to the goals are such as product properties, software release, verification results, duration, effort and cost. Relatively, sixteen metrics were specified for measuring the detected attributes. The study is in two parts of reviewing the state-of-art and a case study at Volvo Cars Corporation. To study the development process and software product, Software Configuration Management (SCM) activities according to the Software Engineering Body of Knowledge (SWEBOK) was applied. Furthermore, an analysis method was used for categorizing the collected data in case-study in order to derive metrics and propose guidelines. Metrics and guidelines were defined based on research result from interviews, observations, pertinent documents and literature reviews. Finally, the validity of proposed metrics was confirmed by experts and manager who take strategical decision in organization.