There is a general agreement on the fact that, at an architectural design stage, the consideration of quality issues are crucial for achieving the overall system quality goals.
The objective of this work is to present and discuss the experience obtained in applying the ABAS (Attribute-Based Architectural Style) technique, for the selection of the architecture of a stock exchange monitoring system.
Non-functional requirements for this application are high availability, platforms heterogeneity, distribution of clients, strict response time. The ABAS technique has been extended introducing the ISO 9126 quality model, to refine the initial high-level quality characteristic, considered relevant to the style, into measurable attributes. In this work we have formalized the definition of architectural styles with quality attributes. For the architecture of the data component, we considered the architectural style for indirect communication and component decoupling, favoring the availability attribute. Several patterns have been studied and evaluated with respect to the availability attribute: Publisher/Subscriber, Repository, Mediator and Broker.
The lessons learned were that the ABAS technique extended with the ISO 9126 quality model and the formal definition of the style, was useful to guide the selection of the architectural pattern. It helped to organize the specialist knowledge and to reason about the relevant quality characteristic of the pattern.
Nicole Levy is professor at the University of Versailles, Saint Quentin en Yvelines. She leads a group on the design of software architecture based on the formalization of the development.
She collaborates since several years with F. Losavio, Professor at the Central University of Venezuela.
Previously she has been assistant professor in Nancy, France, working in the LORIA research center. She participated to the development of Proplane, a formal specification construction model independent of the language of expression, guiding the specifier in its work and memorising his/her expertise.