Comparing Time-Triggered Ethernet with Till Steinbach, Franz Korf, - - PowerPoint PPT Presentation

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Comparing Time-Triggered Ethernet with FlexRay:

An Evaluation of Competing Approaches to Real-time for In-Vehicle Networks Till Steinbach {till.steinbach, korf, schmidt} @informatik.haw-hamburg.de

Hamburg University of Applied Sciences

May, 20th 2010

1 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Introduction

Motivation and Problem Statement

FlexRay is considered the next generation technology for in-car networks

2 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Introduction

Motivation and Problem Statement

FlexRay is considered the next generation technology for in-car networks Bandwidth requirements increase rapidly Current in-vehicle networks are inhomogeneous

2 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Introduction

Motivation and Problem Statement

FlexRay is considered the next generation technology for in-car networks Bandwidth requirements increase rapidly Current in-vehicle networks are inhomogeneous Time-triggered Ethernet promises real-time and best-effort traffic in one backbone network

2 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Introduction

Motivation and Problem Statement

FlexRay is considered the next generation technology for in-car networks Bandwidth requirements increase rapidly Current in-vehicle networks are inhomogeneous Time-triggered Ethernet promises real-time and best-effort traffic in one backbone network Usage of components of the shelf Benefit from the expertise of the plenty Ethernet developers

2 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Introduction

Objectives

Competitive analysis of FlexRay and TTEthernet Show eligibility of TTEthernet for in-vehicle applications Provide framework for performance appraisals of Ethernet based in-vehicle communication

3 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Background

FlexRay Internals

static segment dynamic segment symbol window network idle time bus cycle i bus cycle i+1 static segment

ECU ECU ECU Time-Triggered Messages Event-Triggered Messages Bandwidth 10MBit/s per channel 4 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Background

TTEthernet

TTE-Switch Infotainment chassis LIN-Gateway

BE BE BE

t

RC

t

TT

Time-Triggered Message

RC BE

Rate-Constrained Message Best-Effort Message ECU LIN-BUS

TT TT

t cycle

TT TT RC BE BE BE

t cycle Linkspeed 100MBit/s or 1GBit/s 5 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Comparison

Framework and Parameters

Analytical framework that can be adapted to varying topologies Concrete results based on a sample configuration

Topology with two stars / switches 16 ms cycle

Latency and Jitter are calculated for the sample configuration Bandwidth is compared over various payload sizes

6 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Results

Latency and Jitter

The analytical results for jitter and latency of FlexRay and TTEthernet are comparable Switching delays in TTEthernet cause higher latency in large topologies Jitter is nearly unaffected by the number of switches in the network FlexRay TTEthernet latency min. payload 12.2µs 24µs latency max. payload 265.2µs 372µs jitter bounds 6.4µs < 10µs

7 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Results

Real-time slots at payload size

! "!! #!! $!! %!! &!!! &"!! &#!! ! "!! #!! $!! %!! &!!! &"!! !"#$%#&%'()'#%*$+'(,%-*.(/01 !"#$%"&'()#*+, '--.*/+01+* '2$+34"#

8 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Results

Net bandwidth at payload size

200 400 600 800 1000 1200 1400 10 20 30 40 50 60 70 80 90 100 TTEthernet FlexRay Bandwidth [Mbit/s] Payload [Byte]

20 40 60 80 100 1 2 3 4 5 6 7 8 9 10

TTEthernet allows varying payload sizes in the cycle and parallel message transmission

9 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Conclusion

FlexRay real-time traffic can be embedded in time-triggered Ethernet The TTEthernet correspondent of a fully utilised FlexRay configuration is utilised by approx 11% Especially for larger packages the bandwidth gain in time-triggered Ethernet is high Bandwidth utilisation can further profit from group communication

10 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Outlook

Currently we build a simulation framework for Ethernet based in-vehicle networks Future work will analyse event-triggered communication, segmentation and priority functionalities Further questions concern consolidation strategies for current in-vehicle bus systems.

11 / 12

Comparing Time-Triggered Ethernet with FlexRay Till Steinbach, Franz Korf, Thomas C. Schmidt Introduction Background Results Conclusion & Outlook

Thank you for your attention

12 / 12