TSN is not a stand-alone communication technology, but offers a wealth of new features for Ethernet: Enhancements in areas such as real-time capability and communication latency, synchronization, robustness and configuration. In total, TSN currently comprises over a dozen published IEEE standards, some of which are relevant for industrial automation applications. Initiatives have therefore been launched to standardize manufacturer-independent interoperable components and methods for industrial communication networks and the 'Internet of Things' (IoT). This is currently taking place in IEC 60802 and the OPC Foundation Field Level Communication (FLC).
Such an extension of Ethernet is particularly interesting in the field of industrial edge computing, where the boundaries between cloud or IT infrastructure and production systems and machine controls are becoming blurred. TSN makes it possible to combine the flexibility and openness of IT networks with the reliability and determinism of fieldbus protocols using a single technology and even in a single network. This optimizes and simplifies the overall architecture instead of using a specific technology for each application and connecting them via gateways.

Ethernet is not only used in industrial automation as the basis for various fieldbus technologies, but is also used in many industrial applications as unmanaged Ethernet. In such a network, one of the strengths of standard Ethernet comes into play: the 'distributed intelligence' of Ethernet protocols such as Address Learning and Spanning Tree, with which the network learns, diagnoses and configures the end devices, paths and even problems - such as connection errors - in the network without user intervention.

Over time, the IEEE has standardized a whole series of such protocols; numerous manufacturers of Ethernet switches have implemented them, so that an Ethernet network of unmanaged switches offers sufficient performance and robustness for many applications. Users are generally aware of the dynamic properties of Ethernet mechanisms, such as how long it takes for a network to 'learn' the routes required for communication during start-up, or until a cabling error is detected and an alternative route through the network is established. Therefore, no requirements are placed on these mechanisms with regard to guarantees and determinism.

Three use cases of Ethernet

Open communication structures such as OPC UA over TSN are particularly important for networks in the field of industrial edge computing, where developments are leading to greater convergence between the cloud, IT and production systems.

© TTTech Industrial

As TSN offers extensions to Ethernet, such dynamic and configuration-free mechanisms remain fully available in TSN networks. However, without configuration, all those mechanisms that are used, for example, for prioritizing certain applications such as VoIP in a mixed IT network or for traffic policing, intrusion detection and other security mechanisms in a critical IT infrastructure cannot be used sensibly. As soon as the features of unmanaged Ethernet are no longer sufficient in a TSN-capable network, the configuration of additional mechanisms becomes an issue.
Ethernet with the appropriate TSN add-on features is generally suitable for covering the following three types of application, for the most part even without adaptations to the end devices: Flexible and secure IT networks, unmanaged Ethernet use cases, and deterministic fieldbus-type control applications.

When designing and operating such mixed (convergent) network environments, the question now arises as to how the very different network requirements can be configured, expanded and checked correctly and with as little effort as possible. After all, IT administrators use completely different Ethernet mechanisms and methods for setting up, managing and diagnosing Ethernet LANs than machine builders and automation engineers do for configuring and testing Ethernet-based fieldbuses.

The benefits of the configuration

What approach can be used to successfully configure TSN?

Configuration in IT networks is usually a complex undertaking due to the wide range of possible uses for Ethernet and the numerous manufacturer-specific features. The TSN working group of the IEEE has taken this topic into account during standardization and published a whole series of standards for the manufacturer-independent configuration of TSN features, both for the format of the configuration data and for the interfaces with which this configuration data can be transmitted to the network switches. If TSN products such as network switches and configuration tools support these standards, this enables the end user to configure the TSN network in a simple and comprehensively functional, manufacturer-independent and interoperable manner.

TSN configuration includes the configuration of all mechanisms that TSN offers for network requirements in industrial automation. These include clock synchronization according to IEEE 802.1AS(rev), the various stream-based mechanisms for precise diagnosis and control of bandwidth and/or network latency for critical applications according to the extensions of IEEE 802.1Q (including IEEE 802.1Qav, .1Qbv, .1Qci), but also the preemption feature for more efficient use of bandwidth in networks with long low-priority and short high-priority messages.

These features of TSN are particularly suitable for networking applications in edge and IoT architectures. Here, some of the data is control-relevant, real-time-critical or even safety-critical, while the other, usually much larger part of the data communication is used for diagnostic purposes and other less critical tasks. In such a hybrid network, the configuration must also correspond to the different situations:

- For real-time critical applications, the network must offer real-time guarantees.

- Sporadic alarm messages and calibration data require bandwidth guarantees.

- Diagnostic data should be available on demand without affecting other types of traffic.

Ethernet supports the spontaneous integration of new nodes, provided the network is designed for this. For example, it should be possible to connect a video camera to the machine network without having to reconfigure the network or even the application.

The example of the camera shows how important the inherent concept of TSN is, which guarantees the critical communication requirements of some network components even if other participants without TSN-specific features are integrated into the network. Each participant can therefore make the best possible use of the network within the scope of its own possibilities and is supported by the network itself.