SlideShare a Scribd company logo
1 of 21
Download to read offline
Pre-shaping Bursty Transmissions under IEEE802.1Q
as a Simple and Efficient QoS Mechanism
Nicolas NAVET, University of Luxembourg
Jörn MIGGE, RealTime-at-Work (RTaW)
Josetxo VILLANUEVA, Groupe Renault
Marc BOYER, Onera
SAE INTERNATIONAL
Use-cases for Ethernet in vehicles
SAE INTERNATIONAL
Main TSN QoS protocols on top of Ethernet
8 priority levels for streams
Benefits:
standard and simple
efficient at the highest priority
can be used with shaping in
transmission (“pre-shaping”)
Limitations:
 not fine-grained enough to
for all kinds of requirements
starvation at lowest priority
levels with bursty traffic
IEEE802.1Q
AVB / Credit-Based
Shaper (CBS)
TSN / Time-Aware
Shaper (TAS)
Two egress queues shaped +
6 priority levels below
Benefits:
 Perf. guarantee for AVB classes
 No starvation for best-effort
traffic
Limitations:
 Per class (not stream) shaping
 Not for control traffic
 Not flexible enough with
standard configuration (CMI)
TAS defines egress ports’
gate schedule (open/close)
Benefits:
 Strong time constraints can
be met
 Can be combined with AVB
Limitations:
 Hard to configure
 Rely on a global clock
 Task sched. must be tailored
to communication for best perf.
Temporal QoS = managing interfering traffic
Priority-based Traffic Shaping Time-triggered (TT)
SAE INTERNATIONAL
Main TSN QoS protocols on top of Ethernet
8 priority levels for streams
Benefits:
standard and simple
efficient at the highest priority
can be used with shaping in
transmission (“pre-shaping”)
Limitations:
 not fine-grained enough to
for all kinds of requirements
starvation at lowest priority
levels with bursty traffic
IEEE802.1Q
AVB / Credit-Based
Shaper (CBS)
TSN / Time-Aware
Shaper (TAS)
Two egress queues shaped +
6 priority levels below
Benefits:
 Perf. guarantee for AVB classes
 No starvation for best-effort
traffic
Limitations:
 Per class (not stream) shaping
 Not for control traffic
 Not flexible enough with
standard configuration (CMI)
TAS defines egress ports’
gate schedule (open/close)
Benefits:
 Strong time constraints can
be met
 Can be combined with AVB
Limitations:
 Hard to configure
 Rely on a global clock
 Task sched. must be tailored
to communication for best perf.
Temporal QoS = managing interfering traffic
Priority-based Traffic Shaping Time-triggered (TT)
In the picture too
 Frame-preemption (Qbu+3br)
 Asynchronous traffic shaping (Qcr)
 Cyclic Queuing & Forwarding (Qch)
SAE INTERNATIONAL
QoS support in the switches – on each output port
Traffic
Shaping
Priority-
Based
Scheduling
+
Frame
Preemption
Time-Triggered
Transmission
Up to 8 priority level overall [Figure inspired from Ashjaei2017]
SAE INTERNATIONAL
Under IEEE802.1Q – 3rd hop
High-priority streams
Best-effort
streams
High-
priority
streams
AVB SR-A
Best-effort streams
Under AVB/CBS – 3rd hop
Obtained by
simulation
in RTaW-Pegase
SAE INTERNATIONAL
TSN/TAS: coordinating gate scheduling tables
Sending node
Switch #1
Switch #2
White bands = transmission allowed
grey bands = not allowed
SAE INTERNATIONAL 8
Pre-shaping mechanism
SAE INTERNATIONAL
IEEE802.1Q with pre-shaping in transmission
– Pre-shaping = inserting “well-
chosen” minimum distance between
frames of a segmented message
on the sender side only – other
characteristics of traffic unchanged
– Objective is to spread out
transmissions to reduce latencies
of lower priority traffic
– Pre-shaping typically applies to
video streams to improve perf. of
best-effort
The last packet of the segmented
message must be received by the
deadline, typically 16.66ms for
60FPS camera
SAE INTERNATIONAL
Pre-shaping in practice
Setting idle-times by hand is not practical
– “PRESH” algorithm in RTaW-Pegase
automates it
– No need for dedicated HW unlike CBS &
TAS, implemented in SW in end-systems
– Not part of TSN but not forbidden!
– Find priorities and transmission pauses
between frames of segmented messages
such that
– all bursty frames subject to pre-
shaping meet their deadlines,
– while minimizing as much as
possible the latency of frames in
lower priority traffic classes
SAE INTERNATIONAL 11
Case-study
SAE INTERNATIONAL
Renault Ethernet prototype network
4 Cameras - 30 and 60fps 3 control units
3 domain
master
#Nodes 14
#Switches 5
#streams 41
Workload per
link
Min: <1%,
med:11%
max:60%
Link data rates 100Mbit/s and
1Gbit/s (1 link)
SAE INTERNATIONAL
Case-study: 4 types of traffic
With pre-shaping
in transmission
Pre-shaping parameters
for the 8 video streams
SAE INTERNATIONAL
Case-study: IEEE802.1Q priorities
Command & Control (C&C)
Audio Streams
File & data transfer, diag.
Top priority
Second priority
level
Best-effort
Third priority level
Decreasingpriorities
Video Streams
SAE INTERNATIONAL
Verification techniques
Probability
Response time
Simulation
max.
Upper-bound with
schedulability analysis
Q5Q4
(Actual) worst-case
traversal time (WCTT)
Easily observable events Infrequent events
Testbed &
Simulation
Long
Simulation
Schedulability
analysis
Used in
this study
 Long simulation here = 48 hours of driving  350 000 transmissions for 500ms frames
 Metrics: communication latencies, bandwidth usage and buffer occupancies
SAE INTERNATIONAL
Toolset & Techniques
– RTaW-Pegase: modeling / analysis /
configuration of Ethernet TSN (automotive,
avionics, industry) + CAN (FD)
– Developed since 2009 in partnership
with Onera
– Users across several industries, e.g; Daimler Cars, Airbus Helicopters, CNES, ABB
– Worst-case Traversal Time (WCTT) analysis – used for deadline constraints
– Timing-accurate Simulation – used for average & throughput constraints
– Optimization algorithms for setting the parameters of all supported protocols
Techniques used
SAE INTERNATIONAL
IEEE802.1Q with pre-shaping for Video
Average latencies for best-effort streams
IEEE802.1Q
IEEE802.1Q with pre-
shaping
AVB Tight Idle-Slope
Pre-shaping under IEEE802.1Q improves
average latencies for best-effort streams by
54% on average – up to 86% – similar
performance as using AVB custom classes
Best-effort streams only
Deadlines of C&C, Video, Audio met
– like without Pre-shaping
Averagelatencies(ms)
SAE INTERNATIONAL
IEEE802.1Q with pre-shaping for Video
Worst-case latencies for best-effort streams
IEEE802.1Q
IEEE802.1Q with pre-
shaping
AVB Tight Idle-Slope
Pre-shaping under IEEE802.1Q improves
worst-case latencies for best-effort streams
by 66% on average – up to 90% - similar
performance as using AVB custom classes
Worst-caselatencies(ms)
Best-effort streams only
SAE INTERNATIONAL 19
Discussion & conclusion
SAE INTERNATIONAL
Pre-shaping Pros and Cons
– Simple, compatible with standard IEEE801Q HW and as effective as AVB/CBS in our
experiments but
1. No protection against “babbling idiots” unlike CBS and TAS – per stream policing of Qci
could offer a solution
2. Adding frames to the system may require a reconfiguration of all flows subject to pre-
shaping (unlike AVB with standard parameters)
3. Setting pre-shaping parameters requires dedicated tool support
4. As there is no reshaping along a path, efficiency decreases with the number of hops
5. Pre-shaping is an additional specification to ECU suppliers which has a cost for OEMs,
but pre-shaping can be implemented on a subset of nodes only (e.g., 5 out of 14 in
our case-study)
Any questions? Contact us
nicolas.navet@uni.lu
jorn.migge@realtimeatwork.com

More Related Content

More from RealTime-at-Work (RTaW)

Practical Use Cases for Ethernet Redundancy
Practical Use Cases for Ethernet RedundancyPractical Use Cases for Ethernet Redundancy
Practical Use Cases for Ethernet RedundancyRealTime-at-Work (RTaW)
 
Towards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture Design
Towards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture DesignTowards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture Design
Towards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture DesignRealTime-at-Work (RTaW)
 
Early-stage Bottleneck Identification and Removal in TSN Networks
Early-stage Bottleneck Identification and Removal in TSN NetworksEarly-stage Bottleneck Identification and Removal in TSN Networks
Early-stage Bottleneck Identification and Removal in TSN NetworksRealTime-at-Work (RTaW)
 
Early-stage topological and technological choices for TSN-based communication...
Early-stage topological and technological choices for TSN-based communication...Early-stage topological and technological choices for TSN-based communication...
Early-stage topological and technological choices for TSN-based communication...RealTime-at-Work (RTaW)
 
Insights into the performance and configuration of TCP in Automotive Ethernet...
Insights into the performance and configuration of TCP in Automotive Ethernet...Insights into the performance and configuration of TCP in Automotive Ethernet...
Insights into the performance and configuration of TCP in Automotive Ethernet...RealTime-at-Work (RTaW)
 
Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...
Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...
Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...RealTime-at-Work (RTaW)
 
Timing verification of real-time automotive Ethernet networks: what can we ex...
Timing verification of real-time automotive Ethernet networks: what can we ex...Timing verification of real-time automotive Ethernet networks: what can we ex...
Timing verification of real-time automotive Ethernet networks: what can we ex...RealTime-at-Work (RTaW)
 
Insights on the Configuration and Performances of SOME/IP Service Discovery
Insights on the Configuration and Performances of SOME/IP Service DiscoveryInsights on the Configuration and Performances of SOME/IP Service Discovery
Insights on the Configuration and Performances of SOME/IP Service DiscoveryRealTime-at-Work (RTaW)
 
Timing verification of automotive communication architecture using quantile ...
Timing verification of automotive communication  architecture using quantile ...Timing verification of automotive communication  architecture using quantile ...
Timing verification of automotive communication architecture using quantile ...RealTime-at-Work (RTaW)
 
CAN in Automotive Applications: a Look Forward
CAN in Automotive Applications: a Look ForwardCAN in Automotive Applications: a Look Forward
CAN in Automotive Applications: a Look ForwardRealTime-at-Work (RTaW)
 
PEGASE – a robust and efficient tool for worst-case network traversal time ev...
PEGASE – a robust and efficient tool for worst-case network traversal time ev...PEGASE – a robust and efficient tool for worst-case network traversal time ev...
PEGASE – a robust and efficient tool for worst-case network traversal time ev...RealTime-at-Work (RTaW)
 
Automotive communication systems: from dependability to security
Automotive communication systems: from dependability to securityAutomotive communication systems: from dependability to security
Automotive communication systems: from dependability to securityRealTime-at-Work (RTaW)
 
Frame latency evaluation: when simulation and analysis alone are not enough
Frame latency evaluation: when simulation and analysis alone are not enoughFrame latency evaluation: when simulation and analysis alone are not enough
Frame latency evaluation: when simulation and analysis alone are not enoughRealTime-at-Work (RTaW)
 
Virtualization in Automotive Embedded Systems: an Outlook
Virtualization in Automotive Embedded Systems: an OutlookVirtualization in Automotive Embedded Systems: an Outlook
Virtualization in Automotive Embedded Systems: an OutlookRealTime-at-Work (RTaW)
 
In‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look ForwardIn‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look ForwardRealTime-at-Work (RTaW)
 

More from RealTime-at-Work (RTaW) (20)

Practical Use Cases for Ethernet Redundancy
Practical Use Cases for Ethernet RedundancyPractical Use Cases for Ethernet Redundancy
Practical Use Cases for Ethernet Redundancy
 
Towards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture Design
Towards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture DesignTowards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture Design
Towards Computer-Aided, Iterative TSN-and Ethernet-based E/E Architecture Design
 
Early-stage Bottleneck Identification and Removal in TSN Networks
Early-stage Bottleneck Identification and Removal in TSN NetworksEarly-stage Bottleneck Identification and Removal in TSN Networks
Early-stage Bottleneck Identification and Removal in TSN Networks
 
Early-stage topological and technological choices for TSN-based communication...
Early-stage topological and technological choices for TSN-based communication...Early-stage topological and technological choices for TSN-based communication...
Early-stage topological and technological choices for TSN-based communication...
 
Insights into the performance and configuration of TCP in Automotive Ethernet...
Insights into the performance and configuration of TCP in Automotive Ethernet...Insights into the performance and configuration of TCP in Automotive Ethernet...
Insights into the performance and configuration of TCP in Automotive Ethernet...
 
Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...
Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...
Insights on the Performance and Configuration of AVB and TSN in Automotive Ap...
 
Timing verification of real-time automotive Ethernet networks: what can we ex...
Timing verification of real-time automotive Ethernet networks: what can we ex...Timing verification of real-time automotive Ethernet networks: what can we ex...
Timing verification of real-time automotive Ethernet networks: what can we ex...
 
Insights on the Configuration and Performances of SOME/IP Service Discovery
Insights on the Configuration and Performances of SOME/IP Service DiscoveryInsights on the Configuration and Performances of SOME/IP Service Discovery
Insights on the Configuration and Performances of SOME/IP Service Discovery
 
Timing verification of automotive communication architecture using quantile ...
Timing verification of automotive communication  architecture using quantile ...Timing verification of automotive communication  architecture using quantile ...
Timing verification of automotive communication architecture using quantile ...
 
CAN in Automotive Applications: a Look Forward
CAN in Automotive Applications: a Look ForwardCAN in Automotive Applications: a Look Forward
CAN in Automotive Applications: a Look Forward
 
PEGASE – a robust and efficient tool for worst-case network traversal time ev...
PEGASE – a robust and efficient tool for worst-case network traversal time ev...PEGASE – a robust and efficient tool for worst-case network traversal time ev...
PEGASE – a robust and efficient tool for worst-case network traversal time ev...
 
Automotive communication systems: from dependability to security
Automotive communication systems: from dependability to securityAutomotive communication systems: from dependability to security
Automotive communication systems: from dependability to security
 
What fUML can bring to MBSE?
What fUML can bring to MBSE?What fUML can bring to MBSE?
What fUML can bring to MBSE?
 
Multicore scheduling in automotive ECUs
Multicore scheduling in automotive ECUsMulticore scheduling in automotive ECUs
Multicore scheduling in automotive ECUs
 
Frame latency evaluation: when simulation and analysis alone are not enough
Frame latency evaluation: when simulation and analysis alone are not enoughFrame latency evaluation: when simulation and analysis alone are not enough
Frame latency evaluation: when simulation and analysis alone are not enough
 
Virtualization in Automotive Embedded Systems: an Outlook
Virtualization in Automotive Embedded Systems: an OutlookVirtualization in Automotive Embedded Systems: an Outlook
Virtualization in Automotive Embedded Systems: an Outlook
 
Prototypage virtuel à partir de SysML
Prototypage virtuel à partir de SysMLPrototypage virtuel à partir de SysML
Prototypage virtuel à partir de SysML
 
In‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look ForwardIn‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look Forward
 
RTaW-Sim Brochure
RTaW-Sim BrochureRTaW-Sim Brochure
RTaW-Sim Brochure
 
Mécanismes de protection AUTOSAR OS
Mécanismes de protection AUTOSAR OSMécanismes de protection AUTOSAR OS
Mécanismes de protection AUTOSAR OS
 

Recently uploaded

Solar Photovoltaic Plant Project Proposal by Slidesgo.pptx
Solar Photovoltaic Plant Project Proposal by Slidesgo.pptxSolar Photovoltaic Plant Project Proposal by Slidesgo.pptx
Solar Photovoltaic Plant Project Proposal by Slidesgo.pptxAmarHaddad
 
CAMIONES TOYOTA N04C- Engine y HINO 300.
CAMIONES TOYOTA N04C- Engine y HINO 300.CAMIONES TOYOTA N04C- Engine y HINO 300.
CAMIONES TOYOTA N04C- Engine y HINO 300.JavierAlejandroCordo1
 
technical report on EV. EVs can offer benefitssuch as lower operating costs a...
technical report on EV. EVs can offer benefitssuch as lower operating costs a...technical report on EV. EVs can offer benefitssuch as lower operating costs a...
technical report on EV. EVs can offer benefitssuch as lower operating costs a...Bijay Sharma
 
Benefits of Load Planning in Fleet Management
Benefits of Load Planning in Fleet ManagementBenefits of Load Planning in Fleet Management
Benefits of Load Planning in Fleet Managementjennifermiller8137
 
Basic of Firmware & Embedded Software Programming in C
Basic of Firmware & Embedded Software Programming in CBasic of Firmware & Embedded Software Programming in C
Basic of Firmware & Embedded Software Programming in CKapil Thakar
 
Why Won't Your Audi A3 Shift Into Reverse Gear Let's Investigate
Why Won't Your Audi A3 Shift Into Reverse Gear Let's InvestigateWhy Won't Your Audi A3 Shift Into Reverse Gear Let's Investigate
Why Won't Your Audi A3 Shift Into Reverse Gear Let's InvestigateAutowerks
 
Torque converter In automobile engineering
Torque converter In automobile engineeringTorque converter In automobile engineering
Torque converter In automobile engineeringajroy0196
 
Is Your Porsche Cayenne Radio On The Fritz Here's What It Could Mean
Is Your Porsche Cayenne Radio On The Fritz Here's What It Could MeanIs Your Porsche Cayenne Radio On The Fritz Here's What It Could Mean
Is Your Porsche Cayenne Radio On The Fritz Here's What It Could MeanHuron Automotive
 
Automotive Bootloader Complete Guide with UDS Frame Format
Automotive Bootloader Complete Guide with UDS Frame FormatAutomotive Bootloader Complete Guide with UDS Frame Format
Automotive Bootloader Complete Guide with UDS Frame FormatKapil Thakar
 
-VDA-Special-Characteristics Special characteristics.pdf
-VDA-Special-Characteristics Special characteristics.pdf-VDA-Special-Characteristics Special characteristics.pdf
-VDA-Special-Characteristics Special characteristics.pdfBorja ARRIZABALAGA URIARTE
 
Nokia Drone Networks - Customer Presentation - MWC2.pdf
Nokia Drone Networks - Customer Presentation - MWC2.pdfNokia Drone Networks - Customer Presentation - MWC2.pdf
Nokia Drone Networks - Customer Presentation - MWC2.pdfJosFerro15
 
Introduction to Automotive Bootloader | Programming Sequence
Introduction to Automotive Bootloader | Programming SequenceIntroduction to Automotive Bootloader | Programming Sequence
Introduction to Automotive Bootloader | Programming SequenceKapil Thakar
 
What Should BMW Owners Know About Steptronic Transmission Problems
What Should BMW Owners Know About Steptronic Transmission ProblemsWhat Should BMW Owners Know About Steptronic Transmission Problems
What Should BMW Owners Know About Steptronic Transmission ProblemsSchearer's Sales & Service Inc
 
Timer Handling in UDS | S3 Server Timer | P2 and P2 Start Timer
Timer Handling in UDS | S3 Server Timer | P2 and P2 Start TimerTimer Handling in UDS | S3 Server Timer | P2 and P2 Start Timer
Timer Handling in UDS | S3 Server Timer | P2 and P2 Start TimerKapil Thakar
 
Battery Management System Circuit Diagram.pdf
Battery Management System Circuit Diagram.pdfBattery Management System Circuit Diagram.pdf
Battery Management System Circuit Diagram.pdfElectric Vehicle Geek
 
Car Audio Systems Enhancing Your Driving Experience
Car Audio Systems Enhancing Your Driving ExperienceCar Audio Systems Enhancing Your Driving Experience
Car Audio Systems Enhancing Your Driving ExperienceAuto Sparky
 
Car Seat Covers and Seat Protection Guide
Car Seat Covers and Seat Protection GuideCar Seat Covers and Seat Protection Guide
Car Seat Covers and Seat Protection GuideAskXX.com
 
Toyota Yaris service manual Free.pdf Toyota Yaris Service manual
Toyota Yaris service manual Free.pdf  Toyota Yaris Service manualToyota Yaris service manual Free.pdf  Toyota Yaris Service manual
Toyota Yaris service manual Free.pdf Toyota Yaris Service manualAutocarmanuals.com
 
Quicker and better: South Korea’s new high-speed train 'EMU-320'
Quicker and better: South Korea’s new high-speed train 'EMU-320'Quicker and better: South Korea’s new high-speed train 'EMU-320'
Quicker and better: South Korea’s new high-speed train 'EMU-320'Hyundai Motor Group
 
Vina Score and Vin Min for almost all the models 2024
Vina Score and Vin Min for almost all the models 2024Vina Score and Vin Min for almost all the models 2024
Vina Score and Vin Min for almost all the models 2024jipohal318
 

Recently uploaded (20)

Solar Photovoltaic Plant Project Proposal by Slidesgo.pptx
Solar Photovoltaic Plant Project Proposal by Slidesgo.pptxSolar Photovoltaic Plant Project Proposal by Slidesgo.pptx
Solar Photovoltaic Plant Project Proposal by Slidesgo.pptx
 
CAMIONES TOYOTA N04C- Engine y HINO 300.
CAMIONES TOYOTA N04C- Engine y HINO 300.CAMIONES TOYOTA N04C- Engine y HINO 300.
CAMIONES TOYOTA N04C- Engine y HINO 300.
 
technical report on EV. EVs can offer benefitssuch as lower operating costs a...
technical report on EV. EVs can offer benefitssuch as lower operating costs a...technical report on EV. EVs can offer benefitssuch as lower operating costs a...
technical report on EV. EVs can offer benefitssuch as lower operating costs a...
 
Benefits of Load Planning in Fleet Management
Benefits of Load Planning in Fleet ManagementBenefits of Load Planning in Fleet Management
Benefits of Load Planning in Fleet Management
 
Basic of Firmware & Embedded Software Programming in C
Basic of Firmware & Embedded Software Programming in CBasic of Firmware & Embedded Software Programming in C
Basic of Firmware & Embedded Software Programming in C
 
Why Won't Your Audi A3 Shift Into Reverse Gear Let's Investigate
Why Won't Your Audi A3 Shift Into Reverse Gear Let's InvestigateWhy Won't Your Audi A3 Shift Into Reverse Gear Let's Investigate
Why Won't Your Audi A3 Shift Into Reverse Gear Let's Investigate
 
Torque converter In automobile engineering
Torque converter In automobile engineeringTorque converter In automobile engineering
Torque converter In automobile engineering
 
Is Your Porsche Cayenne Radio On The Fritz Here's What It Could Mean
Is Your Porsche Cayenne Radio On The Fritz Here's What It Could MeanIs Your Porsche Cayenne Radio On The Fritz Here's What It Could Mean
Is Your Porsche Cayenne Radio On The Fritz Here's What It Could Mean
 
Automotive Bootloader Complete Guide with UDS Frame Format
Automotive Bootloader Complete Guide with UDS Frame FormatAutomotive Bootloader Complete Guide with UDS Frame Format
Automotive Bootloader Complete Guide with UDS Frame Format
 
-VDA-Special-Characteristics Special characteristics.pdf
-VDA-Special-Characteristics Special characteristics.pdf-VDA-Special-Characteristics Special characteristics.pdf
-VDA-Special-Characteristics Special characteristics.pdf
 
Nokia Drone Networks - Customer Presentation - MWC2.pdf
Nokia Drone Networks - Customer Presentation - MWC2.pdfNokia Drone Networks - Customer Presentation - MWC2.pdf
Nokia Drone Networks - Customer Presentation - MWC2.pdf
 
Introduction to Automotive Bootloader | Programming Sequence
Introduction to Automotive Bootloader | Programming SequenceIntroduction to Automotive Bootloader | Programming Sequence
Introduction to Automotive Bootloader | Programming Sequence
 
What Should BMW Owners Know About Steptronic Transmission Problems
What Should BMW Owners Know About Steptronic Transmission ProblemsWhat Should BMW Owners Know About Steptronic Transmission Problems
What Should BMW Owners Know About Steptronic Transmission Problems
 
Timer Handling in UDS | S3 Server Timer | P2 and P2 Start Timer
Timer Handling in UDS | S3 Server Timer | P2 and P2 Start TimerTimer Handling in UDS | S3 Server Timer | P2 and P2 Start Timer
Timer Handling in UDS | S3 Server Timer | P2 and P2 Start Timer
 
Battery Management System Circuit Diagram.pdf
Battery Management System Circuit Diagram.pdfBattery Management System Circuit Diagram.pdf
Battery Management System Circuit Diagram.pdf
 
Car Audio Systems Enhancing Your Driving Experience
Car Audio Systems Enhancing Your Driving ExperienceCar Audio Systems Enhancing Your Driving Experience
Car Audio Systems Enhancing Your Driving Experience
 
Car Seat Covers and Seat Protection Guide
Car Seat Covers and Seat Protection GuideCar Seat Covers and Seat Protection Guide
Car Seat Covers and Seat Protection Guide
 
Toyota Yaris service manual Free.pdf Toyota Yaris Service manual
Toyota Yaris service manual Free.pdf  Toyota Yaris Service manualToyota Yaris service manual Free.pdf  Toyota Yaris Service manual
Toyota Yaris service manual Free.pdf Toyota Yaris Service manual
 
Quicker and better: South Korea’s new high-speed train 'EMU-320'
Quicker and better: South Korea’s new high-speed train 'EMU-320'Quicker and better: South Korea’s new high-speed train 'EMU-320'
Quicker and better: South Korea’s new high-speed train 'EMU-320'
 
Vina Score and Vin Min for almost all the models 2024
Vina Score and Vin Min for almost all the models 2024Vina Score and Vin Min for almost all the models 2024
Vina Score and Vin Min for almost all the models 2024
 

Pre-shaping Bursty Transmissions under IEEE802.1Q as a Simple and Efficient QoS Mechanism

  • 1. Pre-shaping Bursty Transmissions under IEEE802.1Q as a Simple and Efficient QoS Mechanism Nicolas NAVET, University of Luxembourg Jörn MIGGE, RealTime-at-Work (RTaW) Josetxo VILLANUEVA, Groupe Renault Marc BOYER, Onera
  • 2. SAE INTERNATIONAL Use-cases for Ethernet in vehicles
  • 3. SAE INTERNATIONAL Main TSN QoS protocols on top of Ethernet 8 priority levels for streams Benefits: standard and simple efficient at the highest priority can be used with shaping in transmission (“pre-shaping”) Limitations:  not fine-grained enough to for all kinds of requirements starvation at lowest priority levels with bursty traffic IEEE802.1Q AVB / Credit-Based Shaper (CBS) TSN / Time-Aware Shaper (TAS) Two egress queues shaped + 6 priority levels below Benefits:  Perf. guarantee for AVB classes  No starvation for best-effort traffic Limitations:  Per class (not stream) shaping  Not for control traffic  Not flexible enough with standard configuration (CMI) TAS defines egress ports’ gate schedule (open/close) Benefits:  Strong time constraints can be met  Can be combined with AVB Limitations:  Hard to configure  Rely on a global clock  Task sched. must be tailored to communication for best perf. Temporal QoS = managing interfering traffic Priority-based Traffic Shaping Time-triggered (TT)
  • 4. SAE INTERNATIONAL Main TSN QoS protocols on top of Ethernet 8 priority levels for streams Benefits: standard and simple efficient at the highest priority can be used with shaping in transmission (“pre-shaping”) Limitations:  not fine-grained enough to for all kinds of requirements starvation at lowest priority levels with bursty traffic IEEE802.1Q AVB / Credit-Based Shaper (CBS) TSN / Time-Aware Shaper (TAS) Two egress queues shaped + 6 priority levels below Benefits:  Perf. guarantee for AVB classes  No starvation for best-effort traffic Limitations:  Per class (not stream) shaping  Not for control traffic  Not flexible enough with standard configuration (CMI) TAS defines egress ports’ gate schedule (open/close) Benefits:  Strong time constraints can be met  Can be combined with AVB Limitations:  Hard to configure  Rely on a global clock  Task sched. must be tailored to communication for best perf. Temporal QoS = managing interfering traffic Priority-based Traffic Shaping Time-triggered (TT) In the picture too  Frame-preemption (Qbu+3br)  Asynchronous traffic shaping (Qcr)  Cyclic Queuing & Forwarding (Qch)
  • 5. SAE INTERNATIONAL QoS support in the switches – on each output port Traffic Shaping Priority- Based Scheduling + Frame Preemption Time-Triggered Transmission Up to 8 priority level overall [Figure inspired from Ashjaei2017]
  • 6. SAE INTERNATIONAL Under IEEE802.1Q – 3rd hop High-priority streams Best-effort streams High- priority streams AVB SR-A Best-effort streams Under AVB/CBS – 3rd hop Obtained by simulation in RTaW-Pegase
  • 7. SAE INTERNATIONAL TSN/TAS: coordinating gate scheduling tables Sending node Switch #1 Switch #2 White bands = transmission allowed grey bands = not allowed
  • 9. SAE INTERNATIONAL IEEE802.1Q with pre-shaping in transmission – Pre-shaping = inserting “well- chosen” minimum distance between frames of a segmented message on the sender side only – other characteristics of traffic unchanged – Objective is to spread out transmissions to reduce latencies of lower priority traffic – Pre-shaping typically applies to video streams to improve perf. of best-effort The last packet of the segmented message must be received by the deadline, typically 16.66ms for 60FPS camera
  • 10. SAE INTERNATIONAL Pre-shaping in practice Setting idle-times by hand is not practical – “PRESH” algorithm in RTaW-Pegase automates it – No need for dedicated HW unlike CBS & TAS, implemented in SW in end-systems – Not part of TSN but not forbidden! – Find priorities and transmission pauses between frames of segmented messages such that – all bursty frames subject to pre- shaping meet their deadlines, – while minimizing as much as possible the latency of frames in lower priority traffic classes
  • 12. SAE INTERNATIONAL Renault Ethernet prototype network 4 Cameras - 30 and 60fps 3 control units 3 domain master #Nodes 14 #Switches 5 #streams 41 Workload per link Min: <1%, med:11% max:60% Link data rates 100Mbit/s and 1Gbit/s (1 link)
  • 13. SAE INTERNATIONAL Case-study: 4 types of traffic With pre-shaping in transmission Pre-shaping parameters for the 8 video streams
  • 14. SAE INTERNATIONAL Case-study: IEEE802.1Q priorities Command & Control (C&C) Audio Streams File & data transfer, diag. Top priority Second priority level Best-effort Third priority level Decreasingpriorities Video Streams
  • 15. SAE INTERNATIONAL Verification techniques Probability Response time Simulation max. Upper-bound with schedulability analysis Q5Q4 (Actual) worst-case traversal time (WCTT) Easily observable events Infrequent events Testbed & Simulation Long Simulation Schedulability analysis Used in this study  Long simulation here = 48 hours of driving  350 000 transmissions for 500ms frames  Metrics: communication latencies, bandwidth usage and buffer occupancies
  • 16. SAE INTERNATIONAL Toolset & Techniques – RTaW-Pegase: modeling / analysis / configuration of Ethernet TSN (automotive, avionics, industry) + CAN (FD) – Developed since 2009 in partnership with Onera – Users across several industries, e.g; Daimler Cars, Airbus Helicopters, CNES, ABB – Worst-case Traversal Time (WCTT) analysis – used for deadline constraints – Timing-accurate Simulation – used for average & throughput constraints – Optimization algorithms for setting the parameters of all supported protocols Techniques used
  • 17. SAE INTERNATIONAL IEEE802.1Q with pre-shaping for Video Average latencies for best-effort streams IEEE802.1Q IEEE802.1Q with pre- shaping AVB Tight Idle-Slope Pre-shaping under IEEE802.1Q improves average latencies for best-effort streams by 54% on average – up to 86% – similar performance as using AVB custom classes Best-effort streams only Deadlines of C&C, Video, Audio met – like without Pre-shaping Averagelatencies(ms)
  • 18. SAE INTERNATIONAL IEEE802.1Q with pre-shaping for Video Worst-case latencies for best-effort streams IEEE802.1Q IEEE802.1Q with pre- shaping AVB Tight Idle-Slope Pre-shaping under IEEE802.1Q improves worst-case latencies for best-effort streams by 66% on average – up to 90% - similar performance as using AVB custom classes Worst-caselatencies(ms) Best-effort streams only
  • 20. SAE INTERNATIONAL Pre-shaping Pros and Cons – Simple, compatible with standard IEEE801Q HW and as effective as AVB/CBS in our experiments but 1. No protection against “babbling idiots” unlike CBS and TAS – per stream policing of Qci could offer a solution 2. Adding frames to the system may require a reconfiguration of all flows subject to pre- shaping (unlike AVB with standard parameters) 3. Setting pre-shaping parameters requires dedicated tool support 4. As there is no reshaping along a path, efficiency decreases with the number of hops 5. Pre-shaping is an additional specification to ECU suppliers which has a cost for OEMs, but pre-shaping can be implemented on a subset of nodes only (e.g., 5 out of 14 in our case-study)
  • 21. Any questions? Contact us nicolas.navet@uni.lu jorn.migge@realtimeatwork.com