Will TSN succeed in establishing a standardized solution for industrial communication in the second attempt? A few years ago, there was a lot of hype about real-time Ethernet for industrial communication. At that time, the industry was just entering the heterogeneous fieldbus world and experts believed that a standardized communication solution could be created from a hardware perspective by relying on Ethernet as the transmission medium. It quickly became apparent that the IEEE Ethernet specification did not meet the requirements of the industry. As a result, several proprietary real-time Ethernet variants were established - as was already the case in the fieldbus world. The result:

Today's 'Industrial Ethernet' is not the same as 'Industrial Ethernet'. In addition, as with fieldbuses, there were also various communication protocols that are incompatible with each other. Conclusion: Ultimately, only the cables and connectors were standardized!

Or will we actually have a convergent network based on a consistent, uniform communication standard?

© Hilscher

The exciting question is whether the individual user organizations that want to use TSN in the future will be able to agree on a uniform TSN profile. This is because the individual sub-standards do not necessarily all have to be used. In addition, the implementations could differ in detail, which would lead to incompatibilities in the field. This can only be avoided by using a standardized 'common' TSN profile.

While Profinet International and the CLPA have clear ideas about Profinet@TSN and CC-Link@TSN, the topic of OPC-UA@TSN in particular initially divided the camps. While the 'shaper' faction is convinced that TSN in conjunction with OPC UA from the sensor to the cloud is the right approach, other major players in the automation world only want to use OPC UA with TSN from cell level, i.e. for controller-to-controller communication and above. In the case of controller-to-controller up to the cloud, the ERP and MES levels of the classic automation pyramid merge together.

The focus of Hilscher as a technology supplier is neither only on controller-to-controller communication nor on OPC-UA@TSN. This means that Hilscher will support both CC-Link@TSN, OPC-UA@TSN and Profinet@TSN.

Beacon of hope: The IEC/IEEE working group

The IEC/IEEE 60802 working group is providing a ray of hope in the direction of a standardized TSN profile, as it is attempting to define a standardized TSN profile for the industry. If this succeeds and all the major players in automation technology join in, then the vision of the so-called 'converged' (mixed) network could become a reality. We see three possible scenarios that could occur in the next few years:

1. each user organization brews its own TSN soup and replaces its current real-time Ethernet solution with it. As is the case today, neither the hardware nor the communication protocol will be compatible with each other.
2. all user organizations adhere to the results of the IEC/IEEE 60802 working group, i.e. all solutions can be implemented with the same hardware, but the world remains heterogeneous in terms of communication protocols (software stacks and profiles). This means that the solutions remain communicatively incompatible with each other.
3. agreement is reached on a standardized communication solution consisting of a TSN profile and communication protocol. All devices can then be operated in the same network and communicate in the so-called 'converged' network.

If the user organizations agree on a uniform TSN profile, it could be possible for the classic automation pyramid to dissolve completely and everything would be connected. The right scheduling could then be used to send frames - via paths in the network - from one sensor to the cloud in the other corner of the network without affecting other traffic.

While 4-port switches are established today with Profinet IRT - with Ethercat even only 2-port devices - the advantage of TSN would be that multi-port capable switches can be used. For this reason, established switch manufacturers such as Belden/Hirschmann and Cisco are also very interested in TSN. There would be greater flexibility in terms of topology and wiring; this flexibility could be extended to include the cost-effective line structure using 2-port-based end devices. This means that topologically anything would be possible: stars, trees, rings, comb structures and even lines could be realized. Hilscher will continue to offer 2-port-capable TSN switches using the netX SoC family.

Will gigabit become mandatory?

In addition to the topology, there is another question that is always under discussion in connection with TSN: Will TSN run on Gigabit Ethernet right down to the last sensor, or will Gigabit even become mandatory?

Hilscher has the following experience in this regard: Hilscher products are now installed in many compact and temperature-critical devices using small, energy-saving multi-protocol SoCs. Customers are already reducing the power dissipation of their devices to less than 1 watt. This is necessary in order not to influence application processes in the device. Even if the technological possibilities continue to develop as a result of progress, Gigabit Ethernet technology will still be 2 to 3 times more powerful than today's Fast Ethernet-based devices in a few years' time. This would mean a knock-out criterion for some customer devices. As the market cannot do without these devices, it can be assumed that a Fast Ethernet variant (100 Mbit) will always be established alongside the Gigabit Ethernet variant in order to be able to continue using precisely these devices. - The susceptibility to interference should also not be forgotten: here too, the pros and cons of Fast and Gigabit Ethernet must be weighed up carefully.

Hilscher can already cover Fast Ethernet-based TSN today with its available multiprotocol-capable netX family. Customers who are currently using netX can therefore switch to TSN via a software update. Nevertheless, in times of vision systems - which have high bandwidth requirements - we do see a trend towards Gigabit Ethernet. In other words, many future field devices will be equipped with Gigabit Ethernet. For this reason, Hilscher is already working on expanding its netX family to include a Gigabit Ethernet variant - as before, everything from chips and software to the finished plug-in card will ultimately be available.

However, the ultimately exciting question remains unanswered: Can the user organizations agree on a common denominator? If not, we will experience a third fieldbus war! Each user organization will incorporate its own 'flavour' - such as proprietary synchronization mechanisms, its own scheduling, incompatible frame formats or summation frame procedures - into 'its own' TSN and also set its own language - i.e. its own communication protocol - on top of it.

Author: Sebastian Hilscher is Division Manager Development at Hilscher.