A Unified Approach for Predictability Analysis of Real-Time Systems using UML-based Control Flow Information

This position paper presents an overview of a comprehensive approach to predictability analysis based on UML models that targets (distributed) Real-Time Systems (RTS). It is based on a generic control flow analysis methodology of UML 2.0 Sequence Diagrams (SD), augmented with analysis of timing and distribution information. Among many uses, the approach enables resource usage analysis, load forecasting, and dynamic dependency analysis in time-dependent systems. Since standard UML models and extensions (profiles) are used as input, this approach is widely applicable and can be applied in early phases of software life cycle. As an example use, we present how the resource usage of network traffic can be analyzed in a distributed RTS. One advantage of our approach is a common core for all predictability analysis processes, namely control flow analysis of UML sequence diagrams. This leads to a unified predictability analysis process, and reusability of analysis information, and will likely simplify the implementation of predictability analysis algorithms