Edge computing terms and skills

Six edge computing questions to ask about data collection, networking and control systems.

By Nate Kay, P.E. April 2, 2020

Whether we realize it or not, most people use edge computing daily. A prime example is when we use the speech recognition features of our mobile phones to ask for Siri or Cortana. Since language processing takes a lot of computational power the phone first does some initial processing, lightening the load for the server and streamlining the data going to it. If all processing was performed on the phone – it would tax the phone’s resources.

Processing data in the cloud frees up the user’s phone to perform other tasks and allows companies like Google and Apple to update and improve the software. If a phone did no pre-processing before sending data to the cloud, our networks and servers could become bogged down with data. A similar model makes sense in industrial applications.

Edge devices operate at the edge of a local network and provide the interface between control system(s) on the plant floor and the outside network. They act as a bridge between the control system and cloud servers or remote computers, processing data between the control system and servers. Performing data computations on the edge device reduces the traffic and the processing power required by both the control system and remote servers.

Having an edge device also allows users to update functions on the edge device without disrupting the control system. Edge devices also can provide a “firewall” or “air gap,” isolating controls equipment from the public network, for better security. Edge devices also can buffer data if there is network latency or even a network outage. If this happens, the edge device stores the data until the network connection can be restored.

Six edge computing questions to ask

While it might not be easy to know if an edge device is right for an application, asking these questions can provide clarity:

  • Do I need to collect historical data from the control system?
  • If I collect data, are there benefits to storing this data in a central location? For example, could this data be used for reporting, down time analysis, predictive maintenance or inventory tracking?
  • Does the control system need to interface with the outside network, such as the plant network, business systems or the internet?
  • Does the control system receive information from the outside network? For example inventory, recipe and batching systems often require this.
  • Is there a benefit to adding mobile devices or features such as alarm notification?
  • Can non-essential functions be offloaded from the “mission critical” control system? For example, could functions such as image processing and recipe management be moved to a non-production computer?

If the answer to one or more of the above questions is yes, then an edge device may be a good fit for that application.

Edge device benefits

The benefits of using an edge device can be grouped into the following categories.

Sharing data 

  • The edge device provides the interface which allows the control system to share data with external systems.
  • The edge device can act as a bridge or protocol converter, allowing legacy equipment to interface with other devices and networks.
  • Ethernet IP devices can interface with the external network without having to modify the existing network or change IP addresses.

Improving security

  • The edge device provides an additional security layer between the control system and external network.
  • The edge device can provide a fire wall and air gap to help protect the control system.
  • The edge device can provide security monitoring and access control.

Processing and network improvements

  • Moving non-critical functions to an edge device allows the control system to “focus” on the most important tasks.
  • It frees up more memory and processing power for mission critical functions.
  • Non-critical functions, running on the edge device, can be updated and modified without the disrupting production.
  • Reduce network traffic and mitigate the impact of network disruptions.
  • Basic data processing on the edge device can help reduce network traffic.
  • Buffering data on the edge device can reduce the impact of network issues.
  • The edge device can convert data to ‘lighter weight’ messaging protocols such as MQTT, reducing bandwidth consumption and improving efficiency.

Next steps

After completing the evaluation, users need to ask how to select the best device for the applications. What features and functions are needed? These three edge device core considerations should be front of mind.

  • Functionality: The control system should operate with or without the edge device. When considering which functions to locate in an edge device ask – if the edge device is turned off, would the process still be able to run reliably? The answer should be yes.
  • Security: An edge device should restrict direct access to the control system from the outside network. The edge device can provide a means to isolate the control system while still allowing data to flow in both directions.
  • Performance: With performance, users need to ask the following questions:
    • Is the control system processing large amounts of incoming or outgoing data? If so, consider having the edge device process this data.
    • What data processing is being performed in the cloud or a remote server? If moved to an edge device the amount of data sent across the network will be reduced.

An edge device can be anything from a relatively simple, low-cost device to something as robust as an industrial PC. What makes an edge device is more a matter of where and how it is used than the actual hardware itself.

Everything highlighted are some of the core features of an edge device. However, edge devices can do more. A lot more. They can help a basic application become future-ready. Enhanced features found in edge devices include:

  • Performing logic and math calculations.
  • Acting as an HMI and host screens.
  • Acting as an Ethernet switch and incorporate features found in managed switches and routers. For example, network address translation (NAT).
  • Edge devices facilitate Industrial Internet of Things (IIoT)/Industry 4.0 functionality and allow users to perform protocol conversion. For example, network traffic conversion from Modbus TCP/IP, CIP and Profinet protocols to MQTT.
  • Run apps and APIs that directly interface with software running on remote servers or the cloud.
  • Run an operating system such as Linux or Microsoft Windows, allowing users to install off-the-shelf software.
  • Provide additional firewall and network security features and diagnostics.
  • Run a structured query language (SQL) database.

While the core functionality can be found in most edge devices, the enhanced features will provide room for the application to grow into the future.

Selecting an edge device and determining how it is used will depend on the specific application and the customer’s needs. When an edge device is properly selected and configured, the result will be a control system with improved performance, higher security and greater maintainability. Most importantly, it provides meaningful information available to those who benefit from it the most.

Nate Kay, project manager, MartinCSI. Edited by Chris Vavra, associate editor, Control Engineering, CFE Media and Technology, cvavra@cfemedia.com.


Keywords: edge computing, smart devices

Edge devices act as a bridge between the control system and cloud servers.

Edge devices are useful for applications where data from an outside network needs to be gathered and evaluated.

What makes an edge device is more a matter of where and how it is used than the actual hardware itself.

Consider this 

What applications do you think would benefit most from edge computing and why?

Author Bio: Nate Kay is a senior project engineer at MartinCSI.