A COMPUTER ASSISTED STATE MARKING METHOD FOR EXTRACTING PERFORMANCE MODELS FROM DESIGN MODELS

The process of performance evaluation is a very important task in designing any software system. Finding a way to extract performance models from design models is therefore a key issue in the software performance engineering field. This paper presents a method for providing computer support for extracting Markov chains, one of the important representations for numerical performance models, from a performance annotated UML sequence diagram according to the UML Profile for Schedulability, Performance and Time. The method uses a marking scheme in which the system status will be recorded each time a new step is executed in a scenario presented by a sequence diagram. The key algorithm is reviewed and an example used to show how the derivation can be managed.