Regresar

ARCHITECTURE-BASED SYSTEMS EVALUATION: LESSONS LEARNED

Abstract:

Information Systems are today highly complex and users are demanding ever higher quality levels. Therefore, it is useful to have a minimum understanding of the level of quality expected of those systems. Architecture is one of the components most likely to affect system quality; hence it must be evaluated at the early stages of development in order to guarantee quality attributes in the future system. The objective of this research is to present the lessons learned from applying two architectural evaluation methods to the same development process applied to a case. The methods used were the Software Architecture Design Method (Bosch 2000) and the Architecture Tradeoff Analysis Method-ATAM (Clements et al. 2002). The case study was a Knowledge Management System known as PROYECTOS DID KMS, which was applied to Research Projects conducted by Universidad Simón Bolívar. The mainly conclusion is that the major difference between both methods lies in their effectiveness, based on the dimension of the systems architecture to be evaluated. The Bosch’s Method, combined with the Simulation Technique, is suitable for systems that handle large amounts of components, since this technique enables the behavior of the different quality attributes to be quantitatively observed by subjecting them to changes in a relatively short space of time. On the other hand, ATAM is appropriate for relatively small systems, where stakeholders can reach a consensus and the architect can also carefully and qualitatively compare each of the scenarios identified.