From UML component diagram to an executable model based on petri net

Abstract

With the growing use of UML diagrams for software architecture description and the importance of evaluating non-functional requirement in software architecture level, providing an executable model of these diagrams seems to be significant. But on one hand software designers are not familiar with the non-functional requirements analysis and are unable to analyze such requirement easily and on the other hand non-functional requirements can not be evaluated directly by UML diagrams. Therefore, this evaluation should be added to the diagrams. In order to do that, an executable model of software architecture should be provided. These executable models can be Petri net, queuing network, stochastic process algebra and etc. In this paper, the primary assumption is that software architecture is described by UML diagrams and then the role of component diagram on performance evaluation is studied. Ultimately through algorithm, it is transformed into an executable model, called Petri Net (PN).