Markov Chain-based Performance Evaluation of FlexRay Dynamic Segment

With the launch of the new BMW X5, the FlexRay protocol for in-car communication has found its way to seriesproduced vehicles. The FlexRay protocol supports both deterministic and non-deterministic transmission of data frames. FlexRay data frames that are being transmitted in the non-deterministic dynamic segment might become displaced under adverse circumstances. In the design of a FlexRay network it is important to have sound understanding of any implications of certain design decisions on the performance of the overall network, here specifically the displacements in the dynamic segment. This paper proposes a novel approach to performance analysis of the dynamic segment based on Markov chain transient analysis. The model of the dynamic segment is a two-dimensional discrete-time Markov chain, where the discrete time steps represent minislots of the dynamic segment. Model properties and assumptions are discussed and expressions for calculating performance metrics are provided. Finally, measurements obtained from a real FlexRay network are used to test model assumptions and to validate the accuracy of model output.