5 use cases for edge computing in manufacturing

What is edge computing?

Edge computing is a computing approach in which companies process data in close proximity to its starting point. Doing so can improve latency, among other benefits.

"Manufacturing has really recognized the benefits of edge computing, whether it's to help make them more operationally efficient, save money, reduce errors in their products or processes, or create a digital representation of what they're doing in their factories," said Jennifer Cooke, director of edge strategies at market research company IDC.

The devices that are considered to be part of edge computing can vary.

IDC defines devices that are part of edge computing as any computing resources that are not housed in the cloud or a company's data center, Cooke said. Under this definition, edge computing could include servers located at a manufacturing site, as well as multiaccess edge computing resources in which the hardware is located off-site but still geographically close by.

Cloud computing is usually considered a separate concept from edge computing.

However, Shubho Ghosh, senior managing director at the Washington, D.C.-based business consultancy firm FTI Consulting, also includes locally based cloud resources in his definition of edge computing.

Why manufacturers are moving to the edge

Despite some variations on the definition of edge computing, companies are generally using the technology for the same reason.

Increased speed is the top reason manufacturers are interested in edge computing, said Joseph Fitzgerald, managing director at Boston Consulting Group. For some manufacturing use cases, sending data to the cloud for analysis and then sending information back results in too much latency, delaying needed decisions.

However, cloud computing is still an important component of smart manufacturing and Industry 4.0, Fitzgerald said. For example, manufacturers still use cloud resources for much of their computing needs and may even use cloud in conjunction with their edge computing in some cases. For example, an organization may use cloud to train and fine-tune the algorithms they use in their edge computing devices. Organizations may also process some data on the edge and then store the data in cloud servers for further analysis.

Others see companies pairing edge and cloud as well.

"It [is] really an edge and cloud story," Cooke said. "They're fused together."

5 edge computing use cases in manufacturing

Manufacturers are deploying edge computing to support various use cases, and the approach is helping them achieve potential benefits, such as improved maintenance and faster operations.

Here are some of those use cases.

1. Computer vision and video surveillance

Edge computing can help companies process data from their computer vision and video surveillance.

Many manufacturers use computer vision and video surveillance to monitor for safety issues, such as worker injuries and quality control, and use AI to analyze the video, Cooke said. Supply chain leaders require near-instantaneous information about these types of problems, so many manufacturers are using edge computing for this surveillance because of its speed in processing data.

2. Robots and autonomous vehicles

Robots and autonomous vehicles also require fast processing, making edge computing a good fit for this technology as well.

Robots and autonomous vehicles must react quickly to their surroundings, so latency must be as limited as possible, Ghosh said. The data analysis involved with robots and vehicles must take place close to the devices.

In addition, edge computing enables devices to continue working without internet access, which is an advantage for robots and autonomous vehicles as well.

Many manufacturers want their robotic devices, including autonomous vehicles, to be capable of operating safely if connectivity to the internet is suddenly lost, and edge computing can make that happen, Fitzgerald said.

3. Predictive maintenance

Predictive maintenance produces a lot of data, and edge computing can help process that.

"Many equipment manufacturers are now building sensors into the machines themselves [for predictive maintenance] and putting in software platforms to aggregate [the] data coming not just from an individual site, but from all the sites worldwide," Cooke said.

Some of the analysis of that data is performed in the cloud, but edge computing could process some of that as well, Cooke said.

"It is often more cost-effective to do some of the compute on the edge," she said.

It is more cost-effective because the system is processing less data. For example, a user may look at the average reading on a device over a particular range of time and use edge computing to process that data instead of sending all sensor readings to the cloud. The user may be able to learn what they need to know from the device's average reading instead of needing to look at every single sensor reading.

4. Adaptive processes

Since edge computing is faster, the technology can help make processes more nimble.

Edge computing analysis can enable technology to change processes quickly to respond to changing situations, Fitzgerald said. For example, edge computing analysis may identify an issue with a calibration device and then direct immediate adjustments. Making faster adjustments saves money and improves efficiency.

5. Extra security

Edge computing's ability to function without internet access is a must for some industries with high security.

Manufacturers in industries such as aerospace and defense may have devices without internet access because of their sensitive work, Fitzgerald said. Even if devices aren't connected to the internet, edge computing can still process needed data.

Mary K. Pratt is an award-winning freelance journalist with a focus on covering enterprise IT and cybersecurity management.