Accelerate the transition to Industry 4.0 with Industrial Ethernet connectivity

The Fourth Industrial Revolution is changing the way we make products, thanks to the digitization of manufacturing and processing equipment. We have witnessed the benefits of automation technology over the past few decades, and now advances in data processing, machine learning, and artificial intelligence are further promoting the development of automation systems. Today, automation systems are increasingly connected, enabling data communication, analysis, and interpretation, and enabling auxiliary intelligent decision-making and actions in factory areas. Smart factory initiatives are creating new business value by increasing production, asset utilization, and overall productivity. They use new data streams to achieve flexibility and optimize quality, while reducing energy consumption and reducing waste residues. In addition, cloud-connected intelligent systems make manufacturing environments more efficient by supporting mass customization.

The benefit of Industry 4.0 is the ability to make better decisions using ever-increasing amounts of data. Throughout the automation system, timely acquisition and transmission of data depends on connectivity. To cope with the growing amount of data, network technology as well as manufacturing processes and methods must continue to improve. Smart, connected automation environments require digitally connected systems, machines, robots, etc. to create and share information. The communication methods used by these machines and the factory communication networks deployed are the core of the enterprise and the key to promoting the realization of Industry 4.0.

Figure 1. Cloud infrastructure.

All sensors and actuators throughout the plant (even remote locations) need to be seamlessly connected, but this is not possible using existing infrastructure. If enterprise-level data is needed to provide actionable insights in the future, the challenge is to find a way to enable this unprecedented amount of data to be transmitted without disrupting the communication network. This raises the question of how to design, build, and deploy an industrial communication network that can meet the needs of today's automation environment and tomorrow's virtual factory.

Why Industrial Ethernet ?

Since connectivity is at the core of achieving the vision of Industry 4.0, three conditions must be met for enterprises to truly achieve connectivity. First, higher-level information technology (IT) or enterprise infrastructure must be integrated with the control network of the plant. Second, the various existing networks or production units in the plant must coexist in a collaborative manner to support interoperability. Third, we need to achieve seamless and secure connectivity throughout the process environment, from the process terminal all the way to the enterprise cloud.

To meet these challenges, we need to adopt a foundational network technology that supports the goals of interoperability, scalability, and coverage. Widely deployed Ethernet technology is the ideal solution. It provides high bandwidth, supports rapid commissioning, and can be widely deployed in the IT infrastructure of all manufacturing environments.

Figure 2. The convergence of two major domains: information technology (IT) and operational technology (OT).

However, given the need for real-time operations, standard Ethernet is not a viable solution for industrial control infrastructure. Operational technology (OT) control networks need to ensure that communication messages are delivered to their intended locations on time. This ensures that the task or process at hand is operating correctly.  The TCP /IP protocol used to route traffic does not inherently guarantee this level of deterministic performance. In the same way that standard Ethernet enables file sharing or access to network devices such as printers, Industrial Ethernet allows controllers to access data and send instructions from PLCs to sensors, actuators, and robots throughout the factory floor. The key difference is the impact of delayed or undelivered messages. In non-real-time applications, a slow web page update has little impact, but in a manufacturing environment, it can have a huge impact, from wasted materials to accidental personal injury. For control systems to work properly, messages must be delivered reliably and on time every time.

As a result, Industrial Ethernet has become the technology of choice for control-level operational technology. Seamless connectivity is required not only between IT and higher-level OT networks, but also between the various layers of the factory OT network and the end-node sensors, as shown in Figure 3. Today, connectivity requires complex, power-hungry gateways at the lower levels of the OT network, and converged IT/OT networks at the higher levels of Ethernet. Having an interoperable, factory-wide Ethernet-based automation network eliminates the need for these gateways, thereby simplifying the network itself. In fact, the protocol gateways used to translate and support upper-level connectivity on the OT network are not directly addressable and create isolation in the network. This data isolation limits the ability to share information across the factory. This runs counter to the previously described vision of Industry 4.0, where manufacturers want to collect telemetry data from the OT side to drive analytics and business processes on the IT side.

Figure 3. Automation system pyramid.

Since determinism in packet delivery and timing ensures that control applications meet mandatory requirements, many vendors have begun to offer real-time protocols suitable for OT networks. This has led to solutions that are deterministic but rely on the protocols of each vendor. As a result, a large number of incompatible solutions have emerged, each using a different type of communication protocol running in different manufacturing units and not interoperating with each other. This has led to the long-term problem of data isolation or data silos. What is needed is a solution that allows different manufacturing departments using different protocols to coexist and share the network without degrading control traffic. This solution is time-sensitive networking (TSN) - a vendor-neutral real-time Ethernet standard based on the IEEE 802.1 specification. As the name suggests, TSN is about time. It transforms standard Ethernet communication into communication that provides timing guarantees for mission-critical applications. The standard aims to ensure that information can move from one point to another in a fixed, predictable time. In this way, TSN guarantees timely delivery. In order for communication to be predictable, devices on the network must use the same concept of time. The standard defines a method to transmit specific TSN Ethernet frames according to a schedule while allowing non-TSN frames to be transmitted on a best-effort basis (see Figure 4). In this way, TSN enables real-time and non-real-time traffic to coexist on the same network. Because all devices use the same time, important data can be transmitted with low latency and low jitter at speeds of up to gigabits.

Figure 4. Time-sensitive network characteristics.

The goal is to build converged networks where each protocol can share the wire in a deterministic and reliable way. TSN is a toolbox of standards that provides the required determinism. It represents a transition to a reliable, standardized connectivity architecture that eliminates data isolation through dedicated fieldbuses. This network convergence, in turn, drives the generation of more data by increasing the scalability of the network itself, in bandwidths ranging from 10 Mbps to 1 Gbps and beyond.

It is likely that TSN will be adopted throughout new facilities, but will be phased in to parts of existing plants. For manufacturers of field devices, this means that in the near future, traditional Industrial Ethernet solutions will need to be supported alongside TSN.

Extension to process terminal

Our final, and perhaps most impactful, change is the ability to enable seamless connectivity from the end node to the enterprise cloud in process control applications, as shown in Figure 5. Until now, connectivity to the end has been limited to existing 4 mA to 20 mA or available fieldbus technologies. In many implementations, these are hardwired point-to-point connections, limiting the flexibility to grow and expand the network over time. These non-Ethernet-based field communications face several challenges. First, very limited bandwidth (e.g., 1.2 kbps for HART® with 4 mA to 20 mA) limits the amount and speed of information flow. Second, the limited power input of the instrument itself limits the functionality of the instrument. Finally, the overhead of gateways at the control and IT levels is unsustainable. There is also the challenge of operating within a Zone 0 intrinsically safe environment and trying to leverage existing cable networks to support faster and cheaper commissioning.

Figure 5. Seamless connection from terminal to cloud.

All of this is driving the development of the IEEE 802.3cg-2019™ standard for 10BASE-T1L full-duplex communications. This standard, which was recently ratified, defines specifications for 10 Mbps full-duplex communications over single twisted pair cables up to 1 km long with power. Data will now be available in sensors and transmitted through OT and IT infrastructures as Ethernet packets. No conversion is required (which causes latency, consumes power, and incurs cost overhead). Existing network architectures will change (as shown in Figure 5), with remote I/O units converted to Ethernet field switches. Ethernet commands can now be transmitted between controllers and field instruments via 10BASE-T1L multi-port field switches. Insights gained at field nodes can then be transmitted as Ethernet packets (with higher bandwidth) over the field switching network to the PLC/DCS and ultimately to the cloud.

There are several distinct advantages that drive the transition from traditional fieldbuses to Industrial Ethernet. First, existing cabling infrastructure can be reused (up to 1 km), simplifying deployment and reducing retrofit costs. Second, the available power delivered through the cable to the instrument itself, previously limited to 36 mW (optimal when deployed with 4 mA to 20 mA), can now reach 60 W (depending on the cable) or 500 mW in Zone 0 (intrinsically safe applications). The additional available power now supports instrumentation with terminal intelligence and more features. Combined with the currently available 10 Mbit uplink speed, it is expected that more insights can be provided, thereby achieving ROI efficiency for Industry 4.0.