Industrial Software on the Edge

December 14, 2018

Industrial Software on the Edge

Edge computing – using the Internet of Things to decentralize business by putting intelligence in the field with the work — is an important advance in industrial software. Integrating IoT devices and small computers physically separated but digitally connected often leverages small computers, but it’s big business.

Little, Big

The continuing progress of Moore’s Law has made what once was considered a big computer into a small and inexpensive computer. Single-card computers like the Raspberry Pi and the Arduino can now be had for as little as $10, but they are fully capable computers running standard operating systems. Equipped with network interfaces and WIFI, they can take on significant applications on their own.

As these big computers have become little computers, using these little computers has become big. Being able to buy hundreds of small computers for what a single desktop system used to cost means that a whole new world of applications have opened up.

Living on the edge

Of course, we need a new buzzword for this, and we’ve got one: edge computing. The idea is simply that we’re doing more and more of the work in computers that aren’t in a big building downtown, but that are distributed to where the work is actually done – potentially anywhere in the world.

It’s a similar idea to the Internet of Things and depends on many of the same technologies. The Internet of Things, however, is about intelligent things like refrigerators, doorbells, light switches, even light bulbs. Edge computing is what happens when you move the first stage of managing those things out to small computers nearby. The small computers, interacting with a local environment, are the computers at the edge of the Internet.

Why Industrial Software is Moving to the Edge

Using these new technologies means new software development. As with most decisions about systems architecture for industrial software, the move to edge computing is driven by two factors: the workload being performed, and economics. (When we talk about workloads in software architectures, we’re talking about the number of operations, the amount of data consumed, and the time it takes: workload is the name for characterizing these things to inform a design. And economics, of course, is money.)

As IoT devices become ubiquitous, there will be thousands or even millions of smart sensors, smart controllers, and computer-controlled components spread over the whole world. Trying to manage all these devices from a central location is really impossible — the “fan in” of data from millions of devices will overwhelm any feasible network.

Those devices also need decision-making to implement policy. This is a fancy way of pointing out that if you have network-enabled light switches, you still need a way to use the network to tell the light to turn on.

If those problems can be solved, however, the advantages can be immense. With real-time reaction to changes as they happen, operations can be optimized, failures can be identified and managed quickly.

Putting small but capable computers on the edge provides a way to distribute those policy decisions and operational optimizations.

A third issue is that while the networks are very widely available, they’re not ubiquitous yet. There is no constellation of internet-providing satellites, no fleets of network enabled weather balloons and blimps. There are even still places where cellular service is spotty or completely unavailable. (My house, for instance.)

Smart things do eventually need to report back, and often that last mile is the hardest problem. This is another application for edge computing, where the “edge” is now a mobile device like a laptop or phone that can be physically taken across the last mile to collect data and update devices, and then returned from the uncivilized wilds to report on what it found.

Making the case: Industrial Software on the edge

We at Flint Hills Group (FHG) have done a number of projects using IoT and edge computing to satisfy a customer’s needs. All of them shared the characteristics that drive an edge solution: a business case for intelligence distributed into many IoT devices, a potential for large data collections, and limited connectivity.

Industrial software keeping the water working
Photo by Ivan Bandura on Unsplash

Great Plains Industries Gets Into the Flow

Great Plains Industries (GPI) is known world-wide for their lines of industrial fuel-transfer pumps, fuel meters, and flow meters. Traditionally, data collection has been done using LCD displays, or by physically connecting a device in the field. Manual entry is error prone, and making connections using custom hardware that’s time and, well, expensive custom hardware.

GPI wanted a better solution, using commodity mobile devices — phones and tablets — and connecting with Bluetooth. GPI tried — but was initially unable to complete an industrial software solution despite great efforts within the organization. FHG was called in and designed a new user interface, implemented a greatly advanced mobile app, and developed new firmware version needed to bridge the gap — and succeeded with GPI. With their new solution, GPI was able to ship their new solution world-wide. The results were so good that GPI has been able to leverage the mobile app solution to open whole new markets they could not address in the past.

Edwin Fetzer of Great Plains Industries was pleased with the results.

“Flint Hills Group excelled at understanding and delivering exactly what was asked of them, on time and on budget. Their technical expertise lead to an app that is even better than expected.”

Industrial software can make even a table smarter
Photo by Christopher Burns on Unsplash

Industrial Software on the Edge: Making a Dumb Table Smart

A large midwestern company that builds filtration systems required for OSHA compliance came to us to help them ensure that systems were being used properly, policies were being executed faithfully, and most important, OSHA inspections passed with no problems.

The product was a line of downdraft tables, which are dust and fume collection systems used in industrial welding, soldering, sanding and polishing applications.

The FHG solution was based on Raspberry Pi computers. Each computer was equipped with a touchscreen and connected to the internal WIFI network. It collected run statistics, maintenance performed on the system and monitoring for critical operating temperatures while alerting humans by email and texts.

These small computers converted what were essentially dumb mechanical downdraft tables into integrated smart tables. This meant saving endless hours of logging and record keeping while improving maintenance, safety and compliance.

Photo by Ksenia Kudelkina on Unsplash

Equipment Auctions on the Edge

Heavyworth provides accurate valuations of heavy equipment to buyers, sellers and financial institutions. To do this, Heavyworth sends employees out to the farms and motor pools where the equipment is located. Or was.

In the past, auction teams were sent to examine and document equipment to be appraised. This was expensive and time consuming, and staffing was a limit to the number of auction teams available. Scaling meant more cost.

FHG partnered closely with Heavyworth and implemented a mobile solution that maintains a real-time connection when possible, but otherwise caches information when outside of cellular coverage, then automatically makes the data available to auction sites as soon as the device enters an area with dependable cellular coverage. Now for many items, instead of sending a team to the physical location, Heavyworth’s mobile app allows the needed information to be gathered from a phone, reducing costs and eliminating a troublesome bottleneck. The solution is now infinitely scalable.

Heavyworth Founder Dusty Reynolds said,

“They take such a posture of partnership that we view them as an extension of our company. It’s important to have someone that understands your pains and what you’re working towards.”

NASA image courtesy @NASANewHorizons

Industrial Software on the Edge

Like most things in the software industry, the concepts, and even the applications, of edge computing aren’t really new. NASA, for example, really puts computing on the edge, where probes like the Voyagers, New Horizons, and the Jupiter Juno probe are literal light-hours away. Dealing with network issue in these scenarios is even worse than cell coverage in the Colorado mountains and Kansas wheat fields!

What’s different now is that inexpensive commodity computers, ubiquitous Internet, broad cellular coverage, and computer-connected components can be had for pennies instead of millions. This technology is now available to everyone.

Need a custom IoT solution built or just want to know how much it might cost? Take our short custom software development calculator quiz and find out!

Charlie Martin
Consulting Software Engineer

Charlie Martin is a consulting software engineer and writer in Erie, Colorado, with interests in systems architecture, cloud computing, distributed systems in general and innovative applications of blockchains in particular. He is available for consulting through Flint Hills Group.


Charlie Martin
Consulting Software Engineer

Charlie Martin is a consulting software engineer and writer in Erie, Colorado, with interests in systems architecture, cloud computing, distributed systems in general and innovative applications of blockchains in particular. He is available for consulting through Flint Hills Group.