Which tasks and functions do OPC-UA and TSN (time-sensitive network) protocols fulfil respectively in industrial systems?
Rahman Jamal: TSN and OPC-UA operate at different (and compatible) levels of the communications stack. TSN provides network services to assure time synchronization and deterministic delivery of packets. It does not deal with the payload of the packet.
OPC-UA operates at the application layer. It assures that the data in the payload of the packet is in a common format that both the publisher and subscriber understand. As an analogy, if this were a phone call TSN would assure that a connection could be made between two parties with enough quality that they could understand in real-time what was being said. OPC-UA assures they use the same language so the message is communicated.
Which tasks and functions will be left to be handled by traditional Industrial Ethernet systems and fieldbuses in the future?
Rahman Jamal: This is probably a market adoption questions, not a technical question. TSN is capable to cover both high performance and low cost applications. However, some applications will benefit more from converged connectivity with a standard Ethernet back-end. We expect to see controller to controller communication being the first function that will adopt TSN.
Will TSN be able to handle tasks in real-time with cycle times as low as 31,25 µs (and possibly lower in the future)?
Rahman Jamal: TSN is simply a prioritization and quality of service layer on standard Ethernet, so yes. Transport times for Ethernet packets is much below this level, especially when you consider the support of faster networks at GigE and beyond.
For instance, bit times (time on the wire) for a 64 byte frame at GigE rates is 512ns, phy delay is 1-2 µs, switch delay can be roughly 1 µs.
This means that devices with one switch in-between can exchange data in < 10 µs. TSN can support daisy chain operation but also can be used in a star configuration to minimize the number of switch hops between nodes. The TSN standard will not be a limitation for these types of loop rates. The limitation will be in companies building products to support these loop rates, including the end nodes ability to process and respond to data in a few µs.
The OPC UA protocol can also be transferred via traditional Industrial Ethernet systems, such as via timeslots or tunneling.
Why would one need TSN then at all?
Which advantages does TSN have at all over traditional Industrial Ethernet systems?
To what extent do costs play a role here?
To what extent does the wish to achieve transfer rates of 1 GBit/s and higher play a role here?
Rahman Jamal: It’s not reasonable to paint all traditional Industrial Ethernet technologies with one brush. Many different vendors and organizations have invested to make “Ethernet” work for their applications. In some cases, they were able to use standard unmodified Ethernet but constrain the system configuration to achieve reliability and performance.
In other cases, they heavily modified at the hardware level to achieve performance. What all of these organizations really wanted was to be able to use standard Ethernet while meeting their requirements for performance and reliability. No one wanted to give up forward scalability, or interoperability, or flexibility.
These were necessary trade-offs to make the existing technology work. TSN is the response of the IEEE standards body to add the needed infrastructure to standard Ethernet. The adoption rate of industry will be determined by the needs of new applications which are challenging or impossible with Industrial Ethernet.
Bandwidth is one technical example but the broader need is for convergence of IT with OT. This provides data access that is needed for higher level business decisions, flexible manufacturing, and process optimization.
In the US we often refer to this as the Industrial internet of things, in Germany we call it Industrie 4.0.