Abstract
Design of a system architecture normally refers to a pre-defined goal. We consider the goal as the desired (functional or non-functional) aspect that is achieved by providing functionalities or mechanisms that support it. For example, a security goal is supported by a set of security mechanisms. However, there are often several ways towards (partially or completely) achieving a goal, which we refer to as design alternatives. Each design alternative is composed of a subset of the mechanisms needed. Some mechanisms are more important than others, and it is not obvious to what degree the different design alternatives fulfill the goal. It can moreover often be difficult to see what combinations of the design alternatives would result in maximum coverage of the goal. We propose an approach to modeling and analysis of the goals of a system. The approach supports specification of the goal, specification of the design alternatives, a quantification of the degree of fulfillment, as well as a quantification of the degree of overlap with respect to the goal fulfillment across design alternatives. We also propose a visual representation of the degrees of fulfillment and the degrees of overlap. We have evaluated the approach on a case study, and the initial results indicate its feasibility. This paper presents the approach and the evaluation results. It also summarizes experiences and suggestions for further improvements.