Edge Computing and the Cloud Give Intelligent Machines a Balanced Load
Will edge computing be the doom of the cloud? Or is it simply finding its niche? See what roles both concepts play in the Industrial Internet.
Join the DZone community and get the full member experience.Join For Free
“The End of Cloud Computing.” “The Edge Will Eat The Cloud.” “Edge Computing—The End of Cloud Computing as We Know It.”
Such headlines grab attention but don’t necessarily reflect reality—especially in Industrial Internet of Things (IoT) deployments. To be sure, edge computing is rapidly emerging as a powerful force in turning industrial machines into intelligent machines, but to paraphrase Mark Twain: “The reports of the death of cloud are greatly exaggerated.”
The Tipping Point: Edge Computing Hits Mainstream
We’ve all heard the stats—billions and billions of IoT devices, generating inconceivable amounts of big data volumes, with trillions and trillions of U.S. dollars to be invested in IoT over the next several years. Why? Because industrials have squeezed every ounce of productivity and efficiency out of operations over the past couple of decades, and are now looking to digital strategies to improve production, performance, and profit.
The Industrial Internet of Things (IIoT) represents a world where human intelligence and machine intelligence—what GE calls minds and machines—connect to deliver new value for industrial companies.
In this new landscape, organizations use data, advanced analytics, and machine learning to drive digital industrial transformation. This can lead to reduced maintenance costs, improved asset utilization, and new business model innovations that further monetize industrial machines and the data they create.
Despite the “cloud is dead” headlines, GE believes the cloud is still very important in delivering on the promise of IIoT, powering compute-intense workloads to manage massive amounts of data generated by machines. However, there’s no question that edge computing is quickly becoming a critical factor in the total IIoT equation.
What Is Edge Computing?
The “edge” of a network generally refers to technology located adjacent to the machine which you are analyzing or actuating, such as a gas turbine, a jet engine, or magnetic resonance (MR) scanner.
Until recently, edge computing has been limited to collecting, aggregating, and forwarding data to the cloud. But what if instead of collecting data for transmission to the cloud, industrial companies could turn massive amounts of data into actionable intelligence, available right at the edge? Now they can.
This is not just valuable to industrial organizations, but absolutely essential.
Edge Computing vs. Cloud Computing
Cloud and edge are not at war… it’s not an either/or scenario. Think of your two hands. You go about your day using one or the other or both depending on the task. The same is true in Industrial Internet workloads. If the left hand is edge computing and the right hand is cloud computing, there will be times when the left hand is dominant for a given task, instances where the right hand is dominant, and some cases where both hands are needed together.
Scenarios in which edge computing will take a leading position include things such as low latency, bandwidth, real-time/near real-time actuation, intermittent or no connectivity, etc. Scenarios where cloud will play a more prominent role include compute-heavy tasks, machine learning, digital twins, cross-plant control, etc.
The point is you need both options working in tandem to provide design choices across edge to cloud that best meet business and operational goals.
Edge Computing and Cloud Computing: Balance in Action
Let’s look at a couple of illustrations. In an industrial context, examples of intelligent edge machines abound—pumps, motors, sensors, blowout preventers and more benefit from the growing capabilities of edge computing for real-time analytics and actuation.
Take locomotives. These modern 200-ton digital machines carry more than 200 sensors that can pump one billion instructions per second. Today, applications can not only collect data locally and respond to changes on that data, but they can also perform meaningful localized analytics. GE Transportation’s Evolution Series Tier 4 Locomotive uses on-board edge computing to analyze data and apply algorithms for running smarter and more efficiently. This improves operational costs, safety, and uptime.
Sending all that data created by the locomotive to the cloud for processing, analyzing, and actuation isn’t useful, practical, or cost-effective.
Now let’s switch gears (pun intended) and talk about another mode of transportation—trucking. Here’s an example where edge plays an important yet minor role, while cloud assumes a more dominant position. In this example, the company has 1,000 trucks under management. There are sensors on each truck tracking performance of the vehicle such as engine, transmission, electrical, battery, and more.
But in this case, instead of real-time analytics and actuation on the machine (like our locomotive example), the data is being ingested, then stored and forwarded to the cloud where time series data and analytics are used to track performance of vehicle components. The fleet operator then leverages a fleet management solution for scheduled maintenance and cost analysis. This gives him or her insights such as the cost over time per part type, or the median costs over time, etc. The company can use this data to improve uptime of its vehicles, lower repair costs, and improve the safe operation of the vehicle.
What’s Next in Edge Computing
While edge computing isn’t a new concept, innovation is now beginning to deliver on the promise—unlocking untapped value from the data being created by machines.
Published at DZone with permission of Corey Olfert, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.