Ethernet-AVB/TSN (Audio Video Bridging/Time-Sensitive Networking) and AFDX (Avionics Full DupleX switched Ethernet) are switched Ethernet technologies, which are both candidates for real-time communication in the context of transportation systems. AFDX implements a fixed priority scheduling strategy with two priority levels. Ethernet-AVB/TSN supports a similar fixed priority scheduling with an additional Credit-Based Shaper (CBS) mechanism. Besides, TSN can support time-triggered scheduling strategy. One direct effect of CBS mechanism is to increase the delay of its flows while decreasing the delay of other priority ones. The former effect can be seen as the shaping restriction and the latter effect can be seen as the shaping benefit from CBS. The goal of this paper is to investigate the impact of CBS on different priority flows, especially on the intermediate priority ones, as well as the effect of CBS bandwidth allocation. It is based on a performance comparison of AVB/TSN and AFDX by simulation in an automotive case study. Furthermore, the shaping benefit is modeled based on integral operation from network calculus perspective. Combing with the analysis of shaping restriction and shaping benefit, some configuration suggestions on the setting of CBS bandwidth are given. Results show that the effect of CBS depends on flow loads and CBS configurations. A larger load of high priority flows in AVB tends to a better performance for the intermediate priority flows when compared with AFDX. Shaping benefit can be explained and calculated according to the changing from the permitted maximum burst.
With the increase in the number of Electronic Control Units (ECUs) and future requirements for vehicle functions, two SR (Stream Reservation) traffic classes may not be sufficient to ensure fulfilment of constraints for multiple traffic types with individual timing requirements transmitted in the Ethernet-AVB (Audio Video Bridging) networks. The goal of this paper is to determine the worst-case delay for an additional SR traffic class under the CBS (Credit-Based Shaper) algorithm. Delay evaluation is based on the impact analysis of CBS on different priority flows, particularly depending on when the credits of both SR class A and B drain from the worst-case perspective. More specifically, both the impact of CBS and the evolution trends of credit on different priority class flows are first analyzed from the worst-case perspective. Then, for an additional SR class, two types of worst-case delay models are established with the CBS configuration suggestions. Finally, an approach to calculate the worst-case queuing delay is proposed. Moreover, the worst-case end-to-end delay is determined by the network calculus approach and simulation. Numerical results show that the delay bounds of our models are tighter than those of other models, which is beneficial to the development of Ethernet-AVB for in-vehicle networking.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.