Performance evaluation of impact of state machine transformation and run-time library on a C# application

Abstract

State machines are important behavioral models used in Model-Driven Development (MDD) of software applications. UML models are transformed into code and combined with a run-time library realizing state machine notions. Mapping of state machine concepts, including concurrent behavior issues, can be realized in various ways. We discuss several problems of call and time event processing and their impact of an application performance. In experiments, different solutions were evaluated and quantitatively compared. These variants were applied in refactoring of FXU - a framework for C# code generation and application development based on UML classes and state machine models.