Industrial Edge technology is playing an increasingly significant role in the digital transformation of industrial companies and smart manufacturing, in particular. The growing number of sensors and smart devices generating terabytes of data, along with the need for quick and reliable processing of this data, has motivated development of a common IoT architecture for edge devices communicating with IoT cloud platforms.

In this special report, the Industrial Ethernet Book offers comprehensive coverage and offers the perspective of industry leaders on how Industrial Edge technology and products are shaping factory applications.

Impact on the shop floor

CPU Virtualization and open communication

According to Thomas Berndorfer, a member of the Executive Board at TTTech Industrial, one of the key technologies for the introduction of Industrial Edge applications is CPU virtualization.

“Virtualization allows applications to run side-by-side on the same standard IPC hardware. Running applications as Docker containers or in Virtual Machines at the edge reduces hardware costs, improves resource efficiency, and gives easy access to data straight from the machine,” Berndorfer said.

“Open source hypervisors such as ACRN from the Linux Foundation also provide virtualization for real-time applications. This even enables the integration of soft PLC functionality on an edge computing platform, further improving efficiency and data access.”

He added that open communication is another key element of industrial edge computing. Standards such as OPC UA ensure that information can be shared from sensor to cloud in a uniform way. Industrial applications running at the edge, especially within a virtualized software environment, benefit from being able to access data from machines and send it to the cloud without the need for gateways.

Focus on industrial applications

Berndorfer said that edge computing is unique in the way that it enables powerful applications to run on the shop floor, i.e. close to the machines that are generating data. In this way it is an extension of cloud computing, which also consolidates various software and large amounts of data in one place.

Unlike with cloud computing, running industrial applications at the edge allows data to be processed with lower latencies and with fewer security issues.

This gives users the opportunity to analyze and use data in real-time to achieve even greater levels of production efficiency and higher system uptime. By running analytics and complex applications at the source (the network edge) and only sending selected data to the cloud, edge computing also helps to reduce the need for expensive server hardware and bandwidth.

Technology benefits

Edge computing allows machine data to be collected on the shop floor and processed close to the machine for real-time operation, before selected data is transferred to the cloud for storage and/or further analysis.

“Real-time collection, processing, and storage of data enables faster decision making and better process planning, especially for the complex environment of a “smart factory” that needs to work autonomously and with maximum efficiency,” Berndorfer said. “Edge computing allows manufacturers to gain in-depth insights into how machines are running and performing. It increases response time in case of technical problems and ensures that downtimes are limited thanks to accurate planning and scheduling of predictive maintenance.”

Customers and engineers also have access to real-time data e.g. when servicing a machine. With edge computing, manufacturers can also set up digitalization initiatives in brownfield sites where no or unreliable Internet connectivity is available.

An edge platform allows for multiple functions to be converged on one device, thus decreasing the cost of hardware. It also offers options for the inclusion of legacy applications and devices, e.g. via a web-based management system. With an edge computing platform, software and devices can be updated and new application deployed remotely, saving travel cost and minimizing the potential of mistakes.

Challenges engineers face

“Automation engineers face a myriad of challenges that edge computing helps to address. A big challenge is how to retrofit existing systems for industrial IoT applications,” he added. “This requires gaining access to machine data and running advanced new software solutions without disrupting the existing infrastructure. Edge computing provides a platform to enable retrofitting whilst reducing costs.”

Other challenges are service and maintenance. Especially in a post-COVID world we see the difficulties of relying on on-site maintenance. In coming years, this is likely to become even more inefficient and expensive. By enabling remote service and maintenance, edge computing can reduce travel and engineering costs, as well as enabling new ‘as-a-service’ offerings that replace traditional service level agreements.

Edge computing devices

Security and flexibility fuel possibilities

Kris Dornan, large controllers marketing manager at Rockwell Automation, said that industrial edge computing devices create the foundation that make edge applications possible. These devices replace “white box” PCs on the plant floor by moving your plant-floor compute hardware into your controller cabinet, controller chassis or even the controller itself.

On this foundation, engineers can then deploy an operating system and software applications that use capabilities like digital twins, analytics and machine learning that create new possibilities in your operations.

“An industrial edge computing device can be an industrial computer that sits in the same chassis as your controller, a compute module that sits on the same rack as the controller or an industrial controller with a built-in computer,” Dornan said.

“No matter which option you choose, you get similar outcomes. You get a computing solution that can be housed and secured in the same location as your controller. And by putting your control and compute components in closer proximity to each other, or by combining them into one platform, you can speed up the information flow in your operations,” he added.

Unlimited possibilities

Dornan said that the end result is that the possibilities for what you can do with edge computing devices are nearly unlimited. For example, you can combine them with vision systems to detect anomalies on high-speed conveyance lines, providing a fidelity of timestamp that may not otherwise be possible. You can improve process visibility by creating an application that calculates a process value that otherwise is difficult to measure or where a sensor is susceptible to environmental factors. And you can use “trial and error” simulations to test production changes in a safe, virtual environment before implementing them in your physical operations.

According to Matt Masarik, design software marketing manager at Rockwell Automation, PCs are increasingly being removed and prohibited from the plant floor for security reasons.

“Industrial edge computing devices provide a means of retaining this crucial computing capability, while at the same time strengthening cybersecurity and improving responsiveness and decision making,” he said.

For starters, edge computing devices offer inherent cybersecurity benefits. They can reduce your attack footprint by allowing you to remove not only PCs from the plant floor but also components like cabling and switches that are required to tether the PCs to your systems. Edge computing devices also allow you to use advanced capabilities like voice- or face-recognition software to create more secure access to your assets.

Additionally, the ability to create software applications for an edge device can help you monitor, manage and optimize your operations in new ways. For example, instead of staff sifting through historical performance data and manually making changes to improve your process, your edge application can continually learn your process and optimize it in real time. Or your edge application can compare how your actual performance compares against the ideal state and identify where you’re falling short.

“In addition to addressing cybersecurity challenges, industrial edge computing devices can help address another concern: the changing workforce,” Masarik said.

Many operators and engineers who are accustomed to tethered control rooms and who use ladder logic and function block programming are retiring. And as they retire, they’re being replaced by a new generation of workers who have different talents and skillsets. These workers have different expectations for technology – specifically that it be “smart” and connected to a mobile device. They’re also more familiar with other programming languages, like Python and Java.

“Edge computing devices align with the expectations and skillsets of these workers,” he added. “For example, some edge computing devices include a library of application programming interfaces (APIs) that allow different programming languages to communicate with the control processor. This can allow engineers to use their preferred programming language to code some or maybe even all of a controller’s logic.”

Microservice architectures

Multiple automation apps each performing a specific task

Thomas Haneder, Marketing Manager, Industrial Edge Ecosystem at Siemens said that in industrial edge applications, they see a clear trend towards the use of apps in discrete and process manufacturing.

“Instead of a monolithic software that handles all tasks, the automation application is split into multiple apps, with each app performing a specific task,” Haneder stated. “For example, connectivity apps connect the edge system to existing shop-floor devices such as PLCs or provide northbound connectivity to different cloud environments. Other apps cover data storage or visualize data for the operations engineer.”

According to Haneder, the underlying technology used by these apps is a microservice architecture, e.g. based on the Docker standard. This technology enables solutions where apps are loosely coupled and thus extremely modular. With this trend there will be more and more applications and services that complement each other.

An open ecosystem will emerge where independent app developers can contribute their domain expertise. Edge app users will benefit from this trend as they will be able to choose the best apps for their solution from a variety of offerings.

Unique microservice architecture

Haneder said that, because of the microservice architecture, an overall solution can thus be realized from various individual apps that communicate with each other via a defined interface. The interface can be realized by an API or by messages encoded with a predefined data format, exchanged via an MQTT-based data bus.

“Other apps can use this open interface and extend the functionality of the underlying base app. Therefore, it is possible to realize a modular and manufacturer-independent automation solution. For example, a connectivity app from vendor A can be used to collect data from a specific plant, machine or device, this data can be stored in data storage app from vendor B, and then displayed by a homegrown app based on an open source library. The individual apps are encapsulated in themselves and are subject to their own lifecycle management, which allows individual services to be updated independently of the overall system,” he said.

With this microservice architectural style, apps can be developed by utilizing DevOps practices such as continuous integration and continuous delivery (CI/CD) which results a faster time to market. This is especially relevant if apps are developed in-house.

Potential impact on manufacturing

One major impact he sees will be that the number of available apps and therefore the possible solutions for a specific use cases will increase tremendously. The customer can thus put together a customized solution consisting of apps from different vendors or extend existing applications through in-house development and due to the modularity without affecting the behavior of the overall system.

Apps will be distributed via digital marketplaces and subscription-based business models will emerge more and more. Thus, customers can quickly access ready-made solutions and expand their system as needed.

Applications are completely independent of the hardware used and can therefore be ported to more powerful edge hardware depending on the computing power requirements. Automation solutions can thus be scaled better. Docker technology allows access to a variety of available images that can be modified and used to create homebrew apps, enabling rapid prototyping of automation solutions.

Challenges for engineers

“In the automation environment assets as well the relationship between them must be managed efficiently to meet OEE targets. This task becomes exponentially complex at scale. The reason for that are different software or software versions per facility, legacy automation systems and therefore many different protocols but also the management of user access rights,” Haneder said.

With the usage of apps and the concept of centralized management of applications and devices, introduced in Siemens’ Industrial Edge system, he added that it much easier to implement a SW and HW lifecycle management.

“Security patches and software updates can now be managed and distributed centrally with one click, no more manual installation. This allows an efficient execution of maintenance activities,” he added.

“Furthermore, with the centralized and remote management a machine or plant builder can extend the functionality of its plant, machine event on-the-fly and is able to introduce new digital business models to his customers. System integrated connectivity apps gives the automation engineers the possibility to connect automation systems and devices from different vendors.”

IT technology in factory

Software platforms with portals, use of containers, Node Red and more

Hilscher’s IIoT Business Development Manager, Craig Lentzkow, sees several specific technologies enabling Industrial Edge applications. These include: use of software platforms with portals; use of containers to run applications; Node-RED running in a container; and connectivity of sensors directly to the cloud (bypassing PLCs, IO-Link Masters, etc.).

Hilscher recently introduced the netFIELD portfolio of IoT solutions, which enables centrally deployed workloads to be processed at the edge at scale. Based on end-to-end managed services, netFIELD encompasses four edge-enabling technologies.

1) Use of Software Platforms with Portals to manage Edge Devices and Application Containers from remote locations. Scalability to manage millions of devices. Platforms connect the Edge Gateway O.S. to the Portal. The Portal provides an open environment for the configuration of the system with minimal or no programming required.

2) Use of Containers to run applications provides application portability and adds a layer of security by preventing third-party applications that may contain malware from contaminating the system. Containers can be removed and added via the netFIELD platform. This allows customers to easily implement Edge Computing in the Edge Gateway. Customers can include AI applications for data analysis and SCADA applications for tracking, alarming, and trending of data.

3) Node Red, running in a container, provides for a graphical way of extracting data via the Edge Gateway from controllers and I/O devices on machinery. The data can be real-time information and can be routed, customized and protocol converted without additional programming. Converted data is typically routed to Cloud servers or Edge Gateway Computing services for further analysis by AI software.

4) Connectivity of sensors directly to the cloud is possible with sensorEdge Gateways, which are part of the netFIELD portfolio. The sensorEDGE Gateway offers a simple and secure way to use sensors to monitor critical machine operations. Select sensors can be connected, via IO-Link technology, directly to the cloud by using an “Edge Gateway on a chip” that is embedded in the IO-Link module. No controller, PLC, or IO-Link Master is required.

Industrial application solutions

Lentzkow stated that what is unique in Industrial Edge Computing Solutions is not the technologies themselves, but the application of these known IT technologies to the factory floor. Using these open, standards-based technologies provides customers with greater flexibility and lower costs. The results are solutions with advantages such as:

• Remote management of edge devices
• Graphical tools to process real-time manufacturing data
• Container management technology to improve app portability and security.
• Scalability to manage millions of devices
• Cloud or on-premise-based data storage
• Zero touch on-boarding of edge gateways

Specific benefits that can be realized are:

• Detection of potential failure of machine components often referred to as “Predictive Maintenance”. This allows users to schedule repairs “just-in-time” instead of incurring unplanned downtime that can costs hundreds of thousands of dollars.
• Production process improvement using analytics against real-time process data.
• Use of digital twin technology to compare real process events with virtual optimized processes, which creates the ability to see where improvement in the real-time production process can be implemented.

He added that overall manufacturing benefits included reduced machine downtime, streamlined supply chain, improved product quality, reduced cost of manufacturing, happier customers.

“My observation is that the automation engineers are having to learn what these IT technologies can bring to the manufacturing process and how to implement and use these IT technologies,” Lentzkow said.

“Training is a ‘must have’ for these automation engineers. IIoT systems are primarily information gathering systems and they run parallel to the process control systems. They require different tool sets than those used in the OT environment. Knowing how to program using higher-level code such as Node js, C++ and Python are beneficial in creating customer specific solutions, but not necessarily required.”

Leveraging IT & OT data

Continuous optimization, efficiency and productivity

According to Vishal Prakash, Strategic Product Manager at ProSoft Technology, to explore the Industrial Edge more effectively, it is useful to define Industrial Edge Computing, Applications, and Gateway.

Industrial Edge Computing: an ability to process time-sensitive data in real time, closer to the source of the information to enable better information flow and real-time decision making, which will improve productivity and response times, and provide better insights into the process.

Industrial Edge Applications: a fully formed end user application that leverages IT+ OT infrastructure and data with business analytics for continuous optimization.

Industrial Edge Gateway: a single device with upstream and downstream connectivity and edge computing capabilities. Upstream connectivity can be to cloud or other business enterprise systems. Downstream connectivity is to field devices like PLCs, RTUs, sensors, and other smart instruments using industrial protocols like Modbus, EtherNet/IP, DNP3, IEC61850, OPC UA, and MQTT.

Enabling Industrial Edge apps

Prakash told IEB that the main technologies enabling industrial edge applications are:

• Ubiquitous & reliable connectivity technologies such as LTE, 5G, and Wi-Fi
• Decentralized computing infrastructure and cloud computing
• Containerized applications
• Increased intelligence and processing power with small form factor for field sensors and automation controllers

“Reliable and high-speed connectivity with reduced latency is a key enabler for industrial edge applications as it allows real-time information exchange between devices and processes. This enables continuous optimization and more current information for better decision-making, leading to increased efficiency and productivity,” Prakash stated.

“Decentralized computing infrastructure and cloud computing is another enabler for industrial edge applications as users can off-load data that is not real-time or not as critical to be processed and analyzed in the cloud. This frees up local computing resources for real-time data processing and other actions that can increase productivity and efficiency.”

Containerized applications

Prakash said that containerized applications is a relatively new concept in the OT world, but is quickly becoming a key enabler for industrial edge applications as it allows the user to deploy applications consistently and quickly, ensuring consistent results for the same machine and process in any location.

And if the results are made available in the cloud for further analysis, then it will be easy to analyze the impact of the containerized application.

And finally, significant progress in the field of microprocessors has enabled field instruments like sensors and meters to operate faster and more reliably while continually reducing their form factor.

Automation controllers have become smarter and more powerful. These developments enable industrial edge computing because the user has access to more accurate, full data as quickly as possible.

“Industrial edge applications are not new to the industrial world. A PLC or RTU running a remote pump station that has the ability to turn the pump ON and OFF is a very simple industrial edge application,” he added. “The same can be said for a controller that is running a remote well pad. But, these are simple and linear examples of edge applications.”

Today, an industrial edge application leverages OT and IT infrastructure data for continuous optimization to achieve increased efficiency and productivity. A digital or smart well pad uses edge computing capabilities not just to turn the well pad on and off but to operate the system securely, safely, and proficiently.

A smart well pad likely includes a dedicated PAC (Programmable Automation Controller) that is deterministic to control the process of oil production and a gateway to handle communications to an enterprise SCADA (Supervisory Control and Data Acquisition) and/or to a cloud-based information system.

The communications gateway with edge computing capabilities allows the well pad to optimize productivity by using commercial data points like current market demand, pricing, and the product delivery schedule for optimal production. The deterministic PLC will communicate with smart field instruments like temperature sensors, pressure transmitters, valves, and flow meters to ensure the oil is produced safely.

“So, what is unique with this technology is the asynchronous operation of the industrial edge computing gateway and traditional controller that can share data for optimal efficiency,” he said.
Impact on manufacturing

Prakash concluded that there are several technology benefits of industrial edge computing. One significant benefit is better security. Threats are persistent, and as such, security has to be continuous. With the power of cloud computing and edge computing, processes and devices can be continuously monitored and any threats can be quickly identified, isolated, and notified for action.

Another benefit is latency. Manufacturing processes are looking for continuous optimization. But to achieve this, the processes need to be able to analyze data from multiple sources, manipulate the data, and use the result to increase efficiency. With edge computing capabilities, this can be done locally. Traditionally, all of this data would need to have been sent to a central host to analyze and forward the results – this requires more bandwidth and creates a potential single point of failure and delays. These factors affect the process’ productivity and efficiency.

“Main challenges faced by automation engineers today are increased IT (Information Technology) / OT (Operational Technology) convergence, increased security requirements, the need for continuous optimization, and data collection from smart field devices. Industrial edge computing can help solve these challenges. Let’s look at each of the challenges in more detail,” Prakash said.

“IT/OT convergence is the integration of OT and IT networks to facilitate sharing of data that can be used to increase productivity. Edge computing horsepower can help manufacturing processes and systems understand and use IT and enterprise data successfully.”

Focus on security

Since threats are persistent, security has to be continuous. While IT/OT convergence has many benefits, the increased connectivity between OT and IT networks does increase the risk of nefarious attacks on OT systems. Given that OT’s top KPI (Key Performance Indicator) is to keep the machine running, edge computing plays a critical role in boosting OT’s defense. The capabilities of an edge compute device means that it can be easily integrated with IT security, increasing overall reliability and efficiency.

“Continuous optimization is every manufacturer’s dream. Optimization requires reliable data from all sources and the ability to analyze the data quickly with results being acted upon,” Prakash added. “This requires processing power. Edge compute devices provide a decentralized compute infrastructure that can reduce bandwidth and latency, positively impacting productivity.”

“Smart field devices provide a copious amount of diagnostic data and other information that is very useful for predictive maintenance, enhancing the life of assets, reducing mean time to repair, and other objectives. To capture and store this information, then analyze and act, processing power and space is required. Edge compute devices are built for this.”

Process, organize & analyze

Addressing storage, security and real-time performance

Specific technologies enabling Industrial Edge applications in 2021 and beyond focus on storage and processing power and complete platform solutions to handle large amounts of data, according to Dr. Al Beydoun, President and Executive Director at ODVA.

“Local gateways with enhanced storage and processing power are making industrial edge applications possible today. These gateways will likely evolve into more substantial server like appliances as the amount of data for analysis continues to increase. IT technology makes it possible to create containers to allow applications and data to be separated for various use cases and business function needs through software from Docker, Amazon, Microsoft, Google, and so forth,” Beydoun said.

“Free applications, such as Node-Red, allow for custom flow-based programming to visualize and analyze device/machine performance and/or operations data. Complete platform solutions that can process, organize, and analyze large amounts of data are made available through traditional industrial automation vendors.”

Beydoun added that EtherNet/IP contributes to the efficiency of edge computing by making key network health and device diagnostics easily available. EtherNet/IP offers a Standard Network Diagnostic Assembly that creates a known object address inside a device to make a consistent set of diagnostic information with context quickly accessible. Consistent content in a common location, without having to send numerous messages to different CIP paths within the device, is made available regardless of the device type or vendor.

Free edge applications

Free edge applications provide the ability to directly map to device I/O and diagnostics as well as to create custom algorithms making it possible for end users to solve tough operational challenges in an agile manner and to provide custom visualization to operators and management on the fly at a low cost.

“Platform edge solutions enable larger scale solutions such as data lakes that pool data from all around a plant for more detailed analysis via out of the box tools or data scientists,” Beydoun added. “An example of a device that can generate a massive amount of data and can benefit from fast, local processing is an Automated Guided Vehicle (AGV) that needs to keep up to date with a constantly changing factory floor environment.”

One of the potential impacts on manufacturing of low-cost edge computing that can be programmed simply and flexibly is the ability to avoid the temptation to go back to paper and pencil or Excel spreadsheets when a new operational problem arises. Additionally, more comprehensive edge platforms can provide for machine learning that can identify potential device failures, like a motor starting to break down, before it happens and alert management to costly operational trends such as a compressor running inefficiently, to save money.

Engineering challenges

According to Beydoun, automation engineers face significant challenges with the high cost of storing and analyzing data in the cloud, security concerns with sending confidential data outside of factory walls, and the need to have real time results from algorithms.

Edge computing can help shift a variable cost from the cloud to a fixed cost with an edge gateway or appliance. Security concerns are also minimized by keeping confidential data such as food recipes local. Finally, edge devices that keep the data onsite are less likely to suffer from connection outages or time delays.

Al Presher, Editor, Industrial Ethernet Book