Linking quality attributes and constraints with architectural decisions

Quality attributes and constraints are among the principal drivers in architectural decision making processes. Quality attributes are improved or damaged by architectural decisions, while constraints directly include or exclude parts of the architecture (e.g., logical components or technologies). We may determine the impact of an architectural decision in the software quality, or which parts of the architecture are a ected by a constraint, but the hard problem is to know if we are respecting the quality requirements (requirements over the quality attributes) and the imposed constraints with all the architectural decisions made. Currently, the most usual approach is that architects use their own experience to produce software architectures that comply with the expected quality requirements and imposed constraints, but at the end, especially for crucial decisions, the architect has to deal with complex tradeo s between quality attributes and juggle with possible incompatibilities raised by the imposed constraints. To facilitate this task and make architect's decision making processes more reliable and e ective, in this paper we present the Quark method to guide the architects in the software architecture design. Quark relies on a specialized ontology, Arteon, which is in charge of managing the architectural knowledge. The decisional part of Arteon is also presented in this paper.
Preprint