Worst-case delay analysis of master-slave switched ethernet networks

Authors:
M. Ashjaei;M. Liu;M. Behnam;A. Mifdaoui;L. Almeida;T. Nolte
Affiliations:
Mälardalen University, Västerås, Sweden;Mälardalen University, Västerås, Sweden;Mälardalen University, Västerås, Sweden;University of Toulouse, France;University of Porto, Portugal;Mälardalen University, Västerås, Sweden
Venue:
Proceeings of the 2nd International Workshop on Worst-Case Traversal Time
Year:
2012

Citing 5
Cited 0

EtheReal: A Host-Transparent Real-Time Fast Ethernet Switch

ICNP '98 Proceedings of the Sixth International Conference on Network Protocols
Tight end-to-end per-flow delay bounds in FIFO multiplexing sink-tree networks

Performance Evaluation
A novel approach to scalable CAC for real-time traffic in sink-tree networks with aggregate scheduling

valuetools '06 Proceedings of the 1st international conference on Performance evaluation methodolgies and tools
Hard Real-Time Communication over Multi-hop Switched Ethernet

NAS '08 Proceedings of the 2008 International Conference on Networking, Architecture, and Storage
Network Calculus for the Validation of Automotive Ethernet In-vehicle Network Configurations

CYBERC '11 Proceedings of the 2011 International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery

Quantified Score

Hi-index	0.00

Visualization

Abstract

Switched Ethernet is increasingly used in real-time communication due to its intrinsic features such as micro segmentation and high throughput. However, COTS switches may impose long blocking times due to their FIFO queues and can also experience buffer overflow in outgoing queues due to uncontrolled packets arrival. The FTT-SE protocol uses a Master-Slave technique to overcome the COTS switch limitations in real-time applications. Recently, we extended the protocol for large scale networks and in this paper we present the worst-case delay analysis using the Network Calculus formalism for such a network. Moreover, we assess the end-to-end delay of traffic with simulation concluding that the obtained analytical results present a good match with the observed delays, providing uppers bounds that vary between 0% and 50% above the maximum measured values.