In today’s electronic system-level (ESL) design processes, an early analysis of a system’s communication and nce characteristics is becoming a key challenge. The availability of abstract performance analysis and estimation techniques supports an early determination and dimensioning of target architectures and an early integration of the design into its embedding environment. Different optimization algorithms require information rent levels of abstraction. The paper addresses static analytical as well as dynamic simulation-based performance analysis techniques. The paper gives insights into a flexible transaction-based approach for the capturing of multi-level communication characteristics, their matching to an appropriate data model, and their statistical analysis in terms of optimization algorithms, which is demonstrated by two examples.
Axel G. Braun, Joachim Gerlach, Wolfgang Rosenstie