Evaluating edge as a platform
By standardizing on complete edge platforms, end users are simplifying edge software and hardware delivery, development, deployment, and support, while streamlining operations. See four edge platform building blocks and four concerns.
- Edge computing platform components need to work well together.
- Cybersecurity, hardware, software and communication are critical aspects of an edge computing platform.
- Users need to decide whether assembly or acquisition is better when developing an edge computing platform.
Edge Computing Insights
- The modern edge consists of hardware devices running many types of software running in a secure, but accessible manner for manufacturers.
- Developing an edge platform can be daunting, but choosing whether to build the platform via assembly or as separate components can help.
- A strong understanding of how edge computing can help a company’s operations and understanding its specific benefits can go a long way to giving the project focus.
“Edge” is “leading edge” technology when it comes to industrial automation, but this simple term encompasses many elements. The edge can be thought of as instruments monitoring physical characteristics, controllers automating equipment, processors performing computations, and more. Certainly, the edge involves connectivity between all these elements, and enables Industrial Internet of Things (IIoT) applications.
In the context of industrial automation and IIoT, users should beware of definitions that are too narrow. Users should plot out what to expect from a complete edge platform ahead of implementation. A platform is not just any one piece of hardware, such as an edge computer, but a range of hardware options, a variety of software, and extensive connectivity and networking options. An edge platform should provide compatible and extensible solutions, with convenient delivery, development, deployment, and support.
While any edge application is created from one or more “things,” it is important to focus not only on what each element does, but how well all elements interoperate. In fact, for the purposes of scalability, it becomes crucial for numerous edge hardware and software components to connect and compute securely. Designers must carefully contemplate a comprehensive edge platform for technical performance and maintainability.
Four edge platform building blocks: Hardware, software, communications, cybersecurity
For decades, the industrial edge – where physical signals were converted back and forth with the digital realm – consisted of programmable logic controllers (PLCs), and then programmable automation controllers (PACs). These were used in conjunction with human-machine interface (HMI) and supervisory control and data acquisition (SCADA) graphics to provide visualization. Here, limited types and quantities of data could be acted upon in ways that supported the operational teams using these systems. While processing power, networking methods, and communication protocols drastically improved the situation over the years, the possibilities remained constrained.
The modern edge is characterized by many hardware devices running many types of software, all of which communicate extensively in a cybersecure, yet accessible, manner (Figure 1). These pieces and parts are arranged to connect operational technology (OT) assets with information technology (IT) resources.
1. Hardware used for an edge application may include a PLC/PAC as in years past. If control is required, a better choice is to use more modern edge controllers. An edge controller operates like a PLC, performing deterministic real-time control, but includes a general-purpose computing system, making it suitable for executing many types of computational, analytical and IIoT tasks.
Each hardware option is suited to specific application needs. Some basic implementations only may require edge gateways for transmitting data while complex computing applications require the power of an industrial PC (IPC).
2. Software enables the chosen hardware to perform its intended control and computing roles. This includes pre-processing data, aggregating it, putting context to the data, storing it, and forwarding information to other systems, usually higher-level on-site or cloud-based resources. Software also is necessary to support local and remote visualization – via HMI and SCADA graphics – in addition to browser-based or mobile displays and dashboards (Figure 2). Software should follow industry standards and provide an open and extensible environment. However, it is equally important for a software environment to be easy to use so deployments and support can succeed.
3. Communications involve hardware and software and possesses unique requirements depending on whether connectivity is to OT or IT assets. Common OT protocols include EtherNet/IP (ODVA), Profinet (PI North America) and Modbus TCP (Modbus Organization), but many other options, including classic serial communications still in use. IT protocols include OPC UA and message queuing telemetry transport (MQTT). Edge devices must be fluent in OT and IT protocols and provide seamless data exchange between the two.
4. Cybersecurity concerns are a natural consequence of expanded connectivity and accessibility. Traditional industrial devices had little or no security provisions. By contrast, modern devices must include the latest cybersecurity technologies, and be easy to maintain and update with patches. Edge device management should require little administrator effort.
Edge platform assembly or acquisition? Four concerns
To varying degrees, each of the elements of hardware, software, communications and cybersecurity can be obtained or assembled à la carte. However, this customized approach introduces risks including extended project durations. The four most common concerns include:
Expending time and effort evaluating prospective products, and confirming they offer the necessary industrial-grade performance.
Confirming connectivity between the hardware products, and the OT and IT resources they must interact with.
Ensuring secure compatibility across applications and selected operating systems
Investigating future scalability and support.
Users, machine builders and system integrators would much rather focus efforts on developing a valuable application, as opposed to investigating the suitability of all the elements underpinning it. For these and other reasons, choosing products from a portfolio of edge solutions – with some offered as ready-to-configure hardware/software combinations – often makes the most sense for edge implementations.
Edge platform versus piecemeal implementation
A complete industrial edge platform provides the essentials for building edge solutions, from the smallest on-machine or in-plant data gathering project, to large-scale production line integrations. This standardization provides developers with the confidence that each project will function properly as it is deployed and will successfully scale up and integrate with future edge initiatives.
For commercial and industrial projects, developers should consider edge platforms with the following five attributes (Figure 3):
Hardware in a scalable and compatible range from small PLCs, to medium/large PACs, to edge controllers, gateways and IPCs.
Real-time deterministic software engine enabling control logic portability between PLCs, PACs and edge controllers.
General-purpose open and extensible IIoT, visualization and analytics software with the ability to run on edge controllers and IPCs.
Support for many critical OT and IT protocols and the ability to add more.
Availability of optimized and pre-licensed packages of edge hardware/software solutions to simplify deployment for common industrial roles, including HMI, edge computing, edge control, edge analytics and SCADA systems.
While the modern industrial edge and IIoT is about openness, the power of a proven platform enabling these types of projects cannot be overstated. An edge platform empowers users to get implementations off the ground and frees up creativity for the project at hand instead of expending it on product research and testing. This has an added benefit of earlier roadblock discovery so developers can address issues, reducing overall project durations and increasing speed to market.
Embracing the complete edge
End users, original equipment manufacturers (OEMs) and the systems integrators who support them are increasingly implementing edge technologies for a wide range of visualization, analytical, and IIoT projects. While there are many options of hardware and software products for implementing these projects, it is a daunting task for developers to ensure the right mix of compatibility, performance and security.
Fortunately, there are choices available from companies offering complete edge platforms dedicated to industrial applications. This alleviates the stress imposed by a custom approach where components are selected and tested piecemeal. By building on a complete edge platform, developers can focus efforts on optimizing applications, while minimizing the delivery, development, deployment and support processes.
Rich Carpenter is general manager, software, for Emerson’s machine automation solutions. Edited by Chris Vavra, web content manager, Control Engineering, CFE Media and Technology, firstname.lastname@example.org.
Keywords: IIoT, edge computing
See additional edge computing stories at https://www.controleng.com/control-systems/embedded-systems-edge-computing/
Will you integrate separate edge devices or use an edge computing platform?