Industry Articles

Building Seamless and Secure Industrial Networks Using TSN

5 days ago by NXP Semiconductors

Moving toward Industry 4.0 is about implementing highly automated and interconnected factories and infrastructure. Learn how time-sensitive networking (TSN) technologies can make this a reality.

The promise of Industry 4.0 involves transforming physical factories and facilities into connected ecosystems that provide unparalleled flexibility and efficiency by allowing better control and planning of automated industrial systems. Meanwhile, cloud computing and machine learning can augment a facility’s throughput and efficiency by allowing engineers to optimize internal processes and quality control further.

All that said, in a world of competing and often incompatible industrial communication standards, these approaches require unifying guidelines and enabling hardware that bridges the gap between industrial equipment beyond the borders of proprietary communication standards.

 

Figure 1. The TSN-capable LS1028A bridge can combine real-time OT motor control and feedback data (blue and green arrows) and best-effort IT traffic (orange arrows) on the same physical Ethernet network. Image used courtesy of NXP

 

In this article, we’ll investigate the peculiarities of information technology (IT) and operational technology (OT) traffic (Figure 1) and their often competing requirements and how new time-sensitive networking (TSN) standards help alleviate problems that often pose a demanding task for industrial network engineers. 

Finally, the article explains how NXP helps bridge the gap between different industrial standards by introducing TSN capabilities in their state-of-the-art crossover microcontrollers (MCUs), applications processors, and network bridges.

 

IT Data and OT Traffic in Industrial Networks

Besides managing competing, proprietary networking standards, engineers also face the challenge of handling different classes of networking traffic—with often competing requirements). 

When transferring IT data—email traffic, for example—a best-effort approach usually suffices, and deterministic response times are negligible. In this domain, the overall data throughput matters more than reliability and guaranteed response times. 

On the other hand, OT traffic describes data generated by industrial equipment—for example, machine control data, sensor information, and actuator movement instructions. This domain requires fixed time delays, low latency, and predictable jitter. However, data throughput is typically not the primary concern when dealing with OT traffic. Instead, reliability is of utmost importance, as delayed messages may not only become obsolete when they arrive late, but missing information may cause severe problems and failures.

The networking requirements typically change dramatically when an industrial facility grows. More equipment generates more data that needs to be managed, and the aforementioned incompatible industrial protocols exacerbate the complexity problem. 

In addition, engineers must keep security, functional safety, reliability, and robustness in mind. Dealing with security, in particular, is another non-trivial task that must never be an afterthought. A single poorly protected or outdated piece of equipment can pose a potential attack surface for a malicious party, potentially leading to catastrophic outcomes, such as the complete standstill of production.

 

Industrial Networking Hierarchy Challenges

Each node within a classical industrial networking structure communicates with others in a given scope, forming a pyramid-shaped hierarchy. However, moving toward distributed smart infrastructure, where nodes contain some form of intelligence and more advanced processing capabilities, requires a flatter, more modern industrial network hierarchy that increases the following:

  • Reliability
  • Resilience
  • Ease of use
  • Communication capabilities 

 

Besides the cost factor, sensor-to-cloud communication imposes another level of complexity to overcome. It requires the network to transmit frames from the very bottom of the hierarchy (the individual sensors) to the top layers (the cloud). This can be seen in Figure 2. 

 

Figure 2. The traditional industrial networking hierarchy looks like a pyramid, and transmitting frames from sensors to the cloud is complex. Image used courtesy of NXP

 

However, the aforementioned mix of technologies makes this move toward flatter hierarchies in real-time industrial networking practically unviable, as each of the often incompatible protocols needs a bridge to exchange data with equipment using other protocols. Adding such gateways increases the overall network complexity, required maintenance effort, and cost and also leads to potential loss of real-time information and decreased precision.

Finally, engineers want to merge OT and IT traffic on a single network in a deterministic, reliable, secure, and robust way (see Figure 1 again). Industry 4.0 aims to solve all these problems and challenges by introducing a single, unified IP paradigm that allows lower edge sensors and actuators to move data to upper control hierarchies even through the cloud while leaving enough bandwidth for best-effort IT traffic.

 

How TSN Bridges the Gap 

It’s important to understand that TSN is not yet another industrial protocol that exists alongside current standard protocols such as Profinet or EtherCAT. TSN is not even a protocol at all, nor is it another proprietary solution. Instead, TSN is a set of IEEE standards that define mechanisms to transmit data over common Ethernet networks while respecting hard real-time constraints. 

Furthermore, TSN aims to create a unifying hardware implementation (Figure 3) that enables tasks such as multi-stream management with hard real-time capabilities, data aggregation, and transmission of larger frames.

The various TSN standards aim to create flatter network topologies that enable edge-to-edge connections in real-time and transmit best-effort traffic over a single line. In summary TSN solves the problems discussed earlier—the problems of enabling sensor-to-cloud frame transmissions without losing intrinsic real-time values.

 

Figure 3. TSN enables engineers to craft a unifying hardware implementation for performing tasks such as multi-stream management with hard real-time capabilities. Image used courtesy of NXP


TSN is a set of standards, a toolbox, where each sub-standard is dedicated to solving specific problems that often occur in industrial networks. The 802.1AS standard, for example, forms the base of TSN. It is based on the precision time protocol (PTP) and is intended for clock synchronization in time-sensitive applications, thus allowing more complex use cases such as time-aware scheduling.

Another important example of a TSN standard is 802.1Qbv, which describes time-aware shaping and enables different classes of traffic (such as IT and OT) to share the same standard Ethernet line. Interested readers can refer to this All About Circuits article, which discusses commonly used TSN standards in greater detail.

TSN defines standards that help manufacturers of industrial equipment and engineers manage all the different needs that often arise in standard real-time networks across various applications. Except for EtherCAT, all major industrial Ethernet protocols are already shifting toward adopting TSN, increasing compatibility between protocols.

 

Driving the Adoption of TSN

To better understand the application of TSN when it comes to specific hardware, we’ll take a look at some examples from NXP. 

Various devices ranging from low-cost, efficient crossover MCUs (for example, the i.MX RT1170) to more powerful devices—such as the i.MX 8M Plus applications processor—already support TSN. The Layerscape LS1028A is one device specifically designed for use in environments requiring TSN support and support for standard industrial real-time networking protocols, such as IEEE 1588, EtherCAT, Profinet, CAN open, and OPC UA.

By natively supporting the most common protocols, the NXP LS1028A (Figure 4) bridges the gap between proprietary, incompatible networking standards and allows engineers to achieve seamless and secure interoperability across existing manufacturing systems. 

 

Figure 4. Layerscape LS1028A block diagram. Image used courtesy of NXP 

 

Furthermore, the LS1028A can act as a bridge between various TSN and non-TSN industrial protocols, thereby allowing engineers to connect an EtherCAT-enabled system. For example, to other equipment that utilizes another protocol, such as Profinet. 

The LS1028A also includes: 

  • TSN-enabled Ethernet controllers for TSN endpoint applications
  • Dual 64-bit Arm Cortex-A72 processors
  • ECC support in the L1 and L2 cache and RAM
  • A dedicated 3D GPU
  • An LCD controller 
  • State-of-the-art security features ensure the confidentiality and integrity of processed data


In the near future an extended range of devices such as the i.MX 9 series of applications processors and the i.MX RT1180 crossover MCU will also support extensive TSN use cases. In particular, the i.MX RT1180 SoC includes a complete networking solution based on a Gb TSN switch that bridges legacy protocols such as EtherCAT or Profinet toward newest generation that are TSN based—for example Profinet CC-D, CC-Link IE TSN, or OPC-UA Pubsub—within a single chip implementation.

Furthermore, the TSN switch supports multiple protocols to bridge communications between existing industrial systems and those of TSN-enabled devices. The first i.MX 9 series devices, the i.MX 93 family, and the i.MX RT1180 crossover MCU are currently sampling, and will launch in 2023.