Life on the edge of the network
Technology is increasingly responsible for social behavior. Words that have existed for eons receive new meanings. Take “cloud,” for example. People are now likely to associate the term with storing data rather than the weather. Despite this new familiarity with an old word, how many of us know or even care what “the cloud” is?
We care about our data and that we can access it whenever we need to. We are beginning to care more about its security. We are also starting to care about which service providers we should trust to provide access to that data through the cloud infrastructure. And all this is true despite our not knowing how the service works or where our data resides.
“Edge” is now a term we use with similar vagueness. Any conversation involving Internet of Things/Industrial Internet of Things (IoT/IIoT) deployments will reference the edge, represented as beyond the cloud at the end of the known network. This means that, unlike the cloud, the edge has a relatable physical presence.
Identifying exactly what the edge is and where you might find it is context-dependent based on your point of view and your use case. Rather than thinking of the edge as a single, physical point, consider it as metaphysical: an idea or a basis for understanding.
There are several ways this idea can manifest:
- Bandwidth at the edge
From a network engineer’s perspective, the edge could mean the last point of resilient connectivity. In this case, data bandwidth may be the defining factor. An application might not need megabits per second of synchronous bandwidth. A reliable asynchronous connection with kilobits per second or less may be adequate for an edge gateway running a small sensor network.
- Processing at the edge
If access to high computation is your metric, the edge may be where you find the last vestige of serious compute power. This might be a network-attached server, multi-protocol router or a high-performance gateway. With most IoT/IIoT applications, sensing and control are the dominant functions; the edge may be the point beyond which the host application becomes an observer rather than exerting influence.
- Power at the edge
Another physical viewpoint is that the edge is the last network end point or node with access to an unconstrained power source. Battery-powered sensors might challenge such thinking, but with today's energy-saving and harvesting techniques, primary cell batteries can now last several years.
The IoT-enabled dairy farm
Farming is one of many industries set to benefit from putting advanced computing at the edge of the network.
Why is edge computing important?
Extending the reach of compute and control beyond traditional networks is the main cause of excitement about the edge. Applications are no longer constrained by access to faceless data or reliant on humans collecting data. To an enterprise, the edge was the walls of the data center. Now, thanks to IoT/IIoT, the walls are coming down. Remote sensors provide real-time information at the point of origin.
Enterprise systems can have access to live data whenever it is needed, from the farthest locations. Live data from the edge can be significant for every use case. Examples include monitoring a patient’s vital signs while they are in their home, measuring soil pH and moisture percentage in a remote field, or the wave height from a buoy in the mid-Atlantic.
Data from the edge doesn’t just benefit commercial organizations; consumers gain remote control of heating, home security, or even just the ability to check in on their cats.
Data collected, actions executed
While the edge has many flavors, broadly speaking edge is where sensing occurs, data are collected, processes controlled, and operational decisions made. Consequently, hardware at the edge varies depending on the type, collection frequency and volume of data.
In its simplest form, the edge could be a compact battery-powered, Bluetooth-connected environmental sensor. An IIoT gateway would be a more sophisticated example and may include wired network connectivity, considerably more compute resources, and multiple analog and digital sensors and actuators. The second example is more complex, but both exemplify edge: capturing data, analyzing it, acting on it and communicating with a host.
The edge can be anywhere. From desolate oceans, remote crop fields, to densely populated cities. The edge could be the expanse of space or under the hood of your vehicle. The edge can yield insights like never before. Connectivity is no longer a barrier to accessing the edge. Wireless communication protocols from Wi-Fi to cellular, Mesh to LPWAN, and low-Earth orbit satellites offer ubiquitous and omnipresent connectivity.
The rise of artificial intelligence at the edge
Ultra-low power yet capable microcontrollers running compact machine-learning algorithms put artificial intelligence (AI) at the edge. These innovations have sharpened the appetite for IoT/IIoT deployment. Edge AI is bringing autonomy to the edge and freedom from constant high-speed connectivity.
For many applications, the rise of the intelligent edge heralds a new era in how value is extracted from data. Rather than shuttle data to the cloud for analysis and processing, applications can make decisions close to data capture. This approach has real benefits, including reduced latency for faster response times, and lowering bandwidth requirements. Privacy and security also improve since there is no direct interaction with the application's host. Edge devices such as sensors and actuators become decentralized and self-contained, working within limits of devolved authority as determined by the host.
The edge is responsible for making an IoT/IIoT deployment tick. Without the edge, there would be no capture, analysis and control. As the edge gains intelligence, industrial and commercial deployments will advance. Not only will the edge undergo significant transformation, but so will the applications that control it. Access to an ongoing stream of sensor data can help refine machine-learning algorithms to identify subtle and nuanced changes in operating parameters, enabling more effective and efficient behavior.
The edge surrounds us. It's near you right now.
Enjoyed this article? Check out our others
