Employing Requirements Metrics for Automating Early Risk Assessment

The most important decisions in software development projects are those taken at the beginning, because they are the most difficult to change later. Especially the assessment of risks is essential in the early stages of development, because the results are a crucial source of information for early decision making. Risks have to be reassessed at frequent intervals, because they might change over time. Currently, typically experts conduct risk assessment and thus it is expensive to repeatedly perform such risk assessments during the course of a project. To improve on this situation, we propose applying metrics to estimate essential information needed for computing product risks. Previous work in this area uses metrics which are based on source code and thus those metrics cannot be used in the early stages of software development. In this paper, a metricbased risk model is presented, which employs requirements metrics to estimate product risks. Thus, our approach brings automation to risk assessment as well as enables the early assessment of product risks in software development projects. Some of the benefits of this early and automated risk assessment are illustrated through an application scenario for risk-based and requirements-based testing.