Three Performance Models at Work: A Software Designer Perspective

The validation of software performance since the very early phases of the lifecycle is a crucial issue in complex software system design. Nowadays in the software development practice, the percentage of time and effort allocated to this task is still too small to avoid performance bugs, which are late to discover and hard to fix. This is due both to the short time to market and to the special skills needed (and often lacking in the development team) to effectively accomplish early performance validation. Software architecture represents a system abstraction that may support the validation and the predictive analysis of system performance. Different notations/languages are available for representing software architectures under a performance viewpoint. In this paper we focus on performance issues of software architectures and we analyze different performance model notations from a software designer perspective. Goal of the paper is to speculate, through a simple case study, on the descriptive power of three largely used performance model notations and their suitability to gain feedback at the architectural design level. This is a preliminary version. The final version will be published in Electronic Notes in Theoretical Computer Science URL: www.elsevier.nl/locate/entcs