What Are the Real-World Applications for Edge Computing?
Edge computing isn’t exactly new. It’s been doing things like managing traffic, helping rail systems run safely and providing computational and processing power in places with low connectivity like oil rigs, isolated military encampments, ships and aircraft for many years.
Instead of relying solely on an endpoint computer to process the data, edge computing allows for data processing at the edge. If necessary, relevant information can be sent to the endpoint. Only sending the relevant processed data saves a lot of bandwidth, especially for systems that generate terabytes of data every day.
IoMT or IoBT
The Internet of Things (IoT) in homes, Industrial Internet of Things (IIoT) in all types of industrial applications and Internet of Military Things (IoMT)/Internet of Battlefield Things (IoBT) share some basic similarities. The military version of IoT uses various sensors to gather data, share it and send it to edge servers to provide real-time updates to other soldiers on the ground and commanders coordinating elsewhere. The defense version of IoT must also meet much higher thresholds for security and signal impenetrability.
Cloud gaming uses edge servers and the cloud to run games that are then essentially streamed on a device-side clients. With so much data being processed for hardware-intensive games at high resolutions, it’s difficult to provide latency-free controls without edge servers to handle processes closer to the gamer.
Society got its first taste of connected cars back in the mid-1990s when early onboard systems were used to dispatch emergency help after accidents. GPS wasn’t added till later iterations of the service. Remote diagnostics and vehicle health reports started becoming common in luxury vehicles in the early to mid-2000s, and Wi-Fi hotspots powered by 4G LTE was fairly accessible to consumers by the mid-2010s.
Future connected cars operating on edge networks might share data with devices inside the vehicle, in nearby vehicles and outside the vehicle. These connected capabilities may play an important role in road safety and coordinated autonomous vehicle navigation in the not-so-distant future.
Estimates suggest an autonomous vehicle can produce between five and 20 terabytes of data every day. The data might include ongoing information updates on road conditions, location, traffic conditions, speed and other vehicles on the road. Autonomous systems need real-time data aggregating and analysis capabilities to operate safely. While some of the data might be sent to manufacturers, car testers or fleet managers, much of that data is processed solely within the vehicle.
Edge Computing Opportunities in Health Care
Machine learning has obvious applications in medical fields, where abnormal data needs to be quickly identified and differentiated from normal reading taken by the countless sensors used in modern health care. Cutting-edge health care facilities are relying more and more on sensors and devices to monitor patient vitals and diagnostics. AI analysis of massive quantities of collected data rely on edge computing technologies for processing power.
Oil rigs in the middle of the Gulf of Mexico aren’t exactly renowned for their connectivity. Many remote oil and gas wells or platforms are equipped with a host of IIoT devices that gather sensor data to detect problems that could manifest as specific sounds, temperature changes, pressure readings or humidity changes. All that data must be processed on the edge to act as a real-time warning for the sake of worker safety and reduced repair costs or equipment and production loss.
Assembly Lines and Manufacturing
Automated assembly lines have been around for a while, but the IIoT sensors and cameras that can monitor operations, detect problems and schedule maintenance are still relatively new. These sometimes-massive webs of IIoT sensors require significant edge computing power to analyze all that input and make optimal use of the data.
Saving Time and Bandwidth
In nearly every application you could potentially accomplish the same thing, albeit much more slowly, by sending information to a data center and then to an end user, but the bandwidth requirements and delays might significantly hinder the analytical value of the data gathered on and near the edge. In some applications, like forward military applications or industrial edge computing needs in the middle of wilderness or massive bodies of water, sending data back to an end user thousands of miles away may simply not be an option.
Edge computing takes many indispensable forms, and it’s not a tool that will be going away anytime soon. If anything, the rapid growth of data competing for limited bandwidth will make edge computing more important in the coming decades.