Mr. Kleineberg, you presented a switch based on TSN (Time-Sensitive Networking) at the Hannover Messe, even though the specification for this standard has not yet been completed. How do you explain this?
Kleineberg: TSN is not 'a' standard, but a collection of different specifications that together describe the further development of Ethernet switching technology. As with all previous IEEE 802 projects, work on the various sub-standards is started at different times in the working group and completed at different times. The majority of these partial specifications can be used independently of the rest. This also explains how we were able to present devices with TSN at Hannover Messe 2016. The 'Time-Aware Scheduler' implemented in these devices in accordance with IEEE 802.1 Qbv has already been approved. This means there is no risk of developing against a moving target with this partial specification. The good thing about this: We were able to demonstrate with this implementation in the trade fair setup that this TSN standard alone already works very effectively and represents significant added value in Industrial Ethernet switches.

Dr. Oliver Kleineberg, Manager Advance Development Industrial IT at Belden: "Many automation manufacturers will be surprised at how quickly TSN is gaining a foothold in the industry."

© Belden

Hirschmann engineers have been working on the TSN standard in the IEEE 802.1 working group since the first discussions began seven years ago. Why is TSN so important to you?
Kleineberg: To date, the real-time Ethernet market in automation has been characterized by different solutions. If Industry 4.0 is to become a reality, this status quo is no longer tenable. The only sensible solution is to rely on a manufacturer-neutral and universally compatible solution for the underlying basic technology. This means that only IEEE 802.3 and 802.1, the 'fathers' of Ethernet itself, can be considered for this. With the audio and video bridging standardization - AVB for short - which began seven years ago, we saw the opportunity to anchor the topic of real-time Ethernet in the IEEE 802. This led to the development of AVB Generation 2, which was renamed TSN in 2011. For us, TSN is the logical further development of the basic Ethernet technology and the next technological leap that will make the future data networks of the IIoT and the factory of the future possible in the first place thanks to its high bandwidths and simultaneous guarantee of different quality of service. In the long term, TSN will become a standard feature of all Industrial Ethernet devices.

Just over a year ago, TSN was not even an issue in factory automation; since then, it has been seen as a possible networking standard in combination with OPC UA right down to the controllers. Will the application field of hard real-time be spared from TSN?
Kleineberg: TSN is not so much a replacement for established industrial Ethernet systems, but rather an extension. For example, it is possible to use TSN as a transmission technology in established ecosystems such as Profinet or Ethernet/IP. Other organizations such as Sercos International are now also discussing the use of TSN as a transmission technology. The OPC Foundation's initiative is particularly noteworthy here. The combination of TSN with OPC UA creates a completely manufacturer-neutral communication system, including the device and communication models in the higher layers. This means that TSN will be successful in all areas of industrial automation in the long term. The mechanisms specified in TSN for background traffic - 'hard' and 'soft' real-time - are able to map the many different traffic patterns in industrial networks today and in the foreseeable future. Due to the high bandwidth and different quality of service as well as time guarantees, this can be done via a single, continuous network.

What cycle and jitter times can TSN actually be expected to achieve?
Kleineberg: TSN can - from the technical design in the specifications - at least achieve what today's established real-time systems can. In the future - with the inevitable further development of Ethernet technology - even more is possible: If bandwidths higher than 1 GBit/s can be realized cost-effectively in industrial networks, then TSN will automatically grow with it. The achievable cycle times ultimately depend on the devices. Not every TSN device will support arbitrarily low cycle times - nor does it have to. The only important thing is that the technology is scalable so that all market requirements can be met and a continuous network can be implemented.

What was the reaction of your customers to the presentation of the first TSN-compatible switch?
Kleineberg: We had hoped for a positive response to the presentation - but we had not expected such a strong reaction as we experienced at the trade fair. We sometimes had a queue at the presentation wall and two or three parallel discussions at the same time - and that lasted almost the entire week of the fair.

Where did this positive response come from?
Kleineberg: I can only speculate, but feedback came from almost all visitors to the trade fair: What was impressive was the fact that TSN was not presented on demo boards or lab samples, but on series devices - Hirschmann brand RSP switches, which were equipped with a TSN software stand for the trade fair. Many visitors only really realized at this point that TSN is no longer just a topic for the future. - Its time has come!

How quickly will TSN establish itself in factory automation?
Kleineberg: The spread depends on many factors, which vary depending on the vertical market: How urgently is the technology needed? When will a specific sub-standard of TSN be available that is urgently needed? How fast are the innovation cycles in this market? Or: How quickly can upgrades be made? How quickly will a prevailing landscape of industry-specific specifications in this market - for example in energy automation - incorporate TSN into its documents? In terms of factory automation, I expect rapid integration, as there will be a great need for this in the coming years due to the transformation of production networks and the established systems are currently unable to meet this demand in line with the market.