TSN4KMU
Offene Routerplattform für eine echtzeitfähige Kommunikation in industriellen Netzwerken mit hoher Bandbreite.
TSN4KMU contribution to TSN/A 2020 virtual conference
Q1: Which fieldbuses do you use most in recent products? (multiple answers allowed)
| Option | Quantity (multiple answers allowed) |
|---|---|
| Option #1: Ethernet/IP | 17 |
| Option #2: ProfiNet | 10 |
| Option #3: Profibus DP | 3 |
| Option #4: EtherCAT | 10 |
| Option #5: Modbus-RTU | 2 |
| Option #6: Modbus-TCP | 4 |
| Option #7: CC-Link | 3 |
| Option #8: PowerLink | 3 |
| Option #9: CAN | 9 |
| Option #10: DeviceNet | 1 |
| Option #11: TSN | 13 |
| Option #12: Others | 7 |
Q2: In which application areas are you working in the most?
| Option | Quantity |
|---|---|
| Option #1: Industrial Automation | 37 |
| Option #2: Automotive | 25 |
| Option #3: Aerospace | 7 |
| Option #4: Defense | 5 |
| Option #5: Telecommunication | 5 |
| Option #6: 5G | 9 |
| Option #7: Others | 9 |
Q&A
How do you envision the TSN-as-a-cable concept integrating with the centralized management model (CUC-CNC)?
TSN-as-a-cable is currently focussed on centrallized management. A new interface is the CNC-CNC interface on slide 9: it shows a standard TSN CNC, controlled interfacing to TSN-over-X CNCs above to pretend the whole TSN network to be a single TSN switch (with the actual network’s latency and jitter)
What is the reason to choose the DSA tag in favor of usual VLAN (Q) or Q-inQ (Double tagged VLAN)?
DSA is well integrated with Linux, which is used on the project’s endpoints. In addition, using VLANs to forward timestamps is not working out correctly, if using standard TSN switches in the network. DSA-tagged frames are just an unknown higher-layer protocol that does not break anything, other than VLANs that are possibly learned and used for traffic management.
DSA might also be used to identify sets of interconnected network segments, which can be done by VLAN instead, but the most important field is and stays the timestamp, which is not usable inside a VLAN tag.
How do you map a TSN frame (up to 1546 Bytes size) to a CAN frame (up to 8 Bytes size)? Do you use frame segmentations? If yes, how do you assign priorities/frame IDs of the CAN frames corresponding to the TSN frame. How are end-to-end latencies determined?
The mapping is the other way around. A single CAN frame or a set of aggregated CAN frames is forwarded in a single TSN frame. Forwarding should be using scheduled timeslots to guarantee the hard CAN timing. Yet, on high TSN network delays, the gateway to CAN segments might need to pretend to be a CAN-CAN-bridge to not create late collisions on the CAN network.