Thiel, Steffen:
A framework to improve the architecture quality of software intensive systems
Duisburg-Essen, 2005
2005Dissertation
WirtschaftswissenschaftenFakultät für Wirtschaftswissenschaften
Titel:
A framework to improve the architecture quality of software intensive systems
Autor*in:
Thiel, Steffen
Erscheinungsort:
Duisburg-Essen
Erscheinungsjahr:
2005
Umfang:
XXI, 299 S. : Ill., graph. Darst.
DuEPublico 1 ID
Signatur der UB:
Notiz:
Duisburg, Essen, Univ., Diss., 2005

Abstract:

Over the past decade, the amount and complexity of software for almost any business sector has increased substantially. Unfortunately, the increased complexity of software in the systems to be built has often lead to a significant mismatch between the planned and the implemented products. One common problem is that system-wide quality attributes such as safety, reliability, performance, and modifiability are not sufficiently considered in software architecture design. Typically, they are addressed in an ad-hoc and unstructured fashion. Since rationales for architectural decisions are frequently missing, risks associated with those decisions can be neither identified, nor mitigated in a systematic way. Consequently, there is a high probability that the resulting software architecture fails to meet business goals and does not allow the building of an adequate system. This work presents QUADRAD, a framework for Quality-Driven Architecture Development. QUADRAD is capable of improving architecture quality for software-intensive systems in a systematic way. It supports the development of architectures that are optimized according to their essential quality requirements. Such architectures permit the building of systems that are better aligned to the principal market needs and business goals. QUADRAD is complemented by the Architecture Exploration Tool (AET), which supports architecture evaluations and helps in documenting the fundamental design decisions of an architecture. QUADRAD has been validated in three industrial projects. For each of these projects the architecture quality could be significantly increased. The results confirm the hypothesis of this work and demonstrate how critical problems in the transition from requirements to architecture design can be mitigated.