The Internet of Things (IoT) may contain 25 billion devices by 2020, according to Gartner. For CIOs, one of the chief technological challenges with the IoT is the timely collection, processing, and analysis of data. In fact, 92 percent of the 203 IoT professionals who participated in a 2015 IoT survey by Dimensional Research said that they couldn’t capture data fast enough.
Typically, IoT data is collected and transmitted to a cloud or data center, where it is processed and analyzed. This approach is reliable, but time consuming. As a result, edge computing is becoming an increasingly popular and faster approach.
In edge computing, sensors, controllers, and other connected devices collect and analyze IoT data themselves, or transmit it to a nearby computing device, such as a server or laptop, for analysis. When this data processing and analysis occurs at the edge of a network, as opposed to a data center or cloud, the data can be immediately analyzed—and put into action.
Edge computing is sometimes mistakenly called fog computing, but the two are different. In fog computing, a single centralized computing device processes data from multiple end points in a network; in edge computing, each device in a network plays its own role in processing data.
Edge computing works at the individual device, fleet, or plant level. Among its benefits are:
- Real-time or near real-time data analysis as the data is analyzed at the local device level, not in a distant data center or cloud;
- Lower operating costs due to the smaller operational and data management expenses of local devices vs. clouds and data centers;
- Reduced network traffic because less data is transmitted from local devices via a network to a data center or cloud, thereby reducing network traffic bottlenecks;
- Improved application performance as apps that don’t tolerate latency can achieve lower latency levels on the edge, as opposed to a faraway cloud or data center.
Thanks to these advantages, many experts expect edge computing to become a mainstream technology during the next five to 10 years.
Edge computing in action
Here are four present-day examples of edge computing in action:
Envision wind turbines
Envision manages a vast network of 20,000 wind turbines. The 3 million sensors installed on these turbines produce up to 20 terabytes of data at a time. With edge computing, Envision has cut its data analysis from 10 minutes to just seconds, enabling it to increase the wind turbines’ production by 15 percent.
Coca-Cola vending machines
Thanks to its edge server, Coca-Cola Freestyle vending machines can quickly dispense more than 100 different combinations of carbonated and noncarbonated drinks. Freestyle machines transmit sales information to Coca-Cola’s Atlanta headquarters on an hourly basis. As a result, Coca-Cola can continuously fine-tune its sales efforts, based on customer preferences.
Among its data-based discoveries: Coca-Cola found out that Caffeine-Free Diet Coke, which had previously been available in less than one percent of its dispensers, is a top five brand during the afternoon. The upshot: The once largely ignored soda is now widely available, boosting Coca-Cola’s overall sales.
GE digital locomotives
A GE Transportation Evolution Series Tier 4 locomotive contains more than 200 sensors. These embedded sensors constantly collect gigabytes of operational data and process more than 1 billion instructions per second while a locomotive is working. Due to the locomotive’s onboard edge computing system, the train is able to apply algorithms in real-time, enabling it to perform at or near its peak performance levels, thereby improving the owner’s bottom line.
Palo Alto’s smart infrastructure
The city of Palo Alto is investing in a host of edge-friendly IoT projects, such as a parking space sensor program that will notify drivers about available parking spaces, thereby reducing traffic congestion and air pollution. Likewise, a new $3 million smart traffic signal project will enable traffic lights to work in sync with connected vehicles, so drivers aren’t forced to sit at empty intersections, waiting for the traffic light to turn green.
Future-minded CIOs should consider edge computing whenever they have an opportunity to improve a function, product, or service with real-time data analysis. Nikhil Chauhan, a director of marketing at GE Capital, urges CIOs and other interested parties to ponder a handful of relevant edge questions, all of which principally ask, “What would you like to achieve with edge computing’s real-time data processing and analytics tomorrow that you are unable to achieve today?”
As always, CIOs who are interested in adopting a new technology, like edge computing, could save themselves plenty of time, effort, and frustration by observing what other companies in their industry are doing with it and ask themselves, “Does this make sense for us?”