How Edge Compute Works and Why It's a
Edge compute is the data processing that takes place at the network edge to decrease latency and reduce demands on cloud compute and data center resources. Edge computing takes place in intelligent devices — right at the location where sensors and other instruments are gathering and processing data — to expedite that processing before devices connect to the Internet of Things (IoT) and send the data on for further use by enterprise applications and personnel.
The primary reason for the growth of edge compute is efficiency. All of that collected data needs to be processed somewhere. And as the volume of IoT data has increased, more and more of the processing is taking place at the edge. Connected devices today are smarter, enabling the ability to program "edge AI" — artificial intelligence at the edge — a growing trend in edge intelligence.
With decades of experience in the rapidly evolving IoT industry, Digi has a complete product offering for optimizing IoT applications with edge compute functionality.
Delivering Only the Important Data
In IoT, massive amounts of data are collected at the edge of the network, but not all of it is useful. On average, most monitoring data tends to be standard “heartbeat” data. If the data isn’t changing significantly, that means things are working well. For example, it wouldn’t make sense to send hours of data to a distant data center, showing that a machine's vital signs haven’t changed.
In the past, companies would send all of their monitoring data into the cloud or to a corporate data center for processing, analysis and storage. As the IoT has grown, however, the volume of data makes this approach impractical. This is where edge compute enters the picture.
Edge compute performs processing close to where the data originates. That can greatly reduce or even eliminate the cost of the bandwidth needed to transmit it to the cloud or the corporate data center. Some applications do need to examine data at the edge. An intelligent or AI-enabled edge compute process can then immediately assess whether the situation demands a response in real time, or send it on to the data center for analysis.
Data collected at the edge falls into roughly three types:
- It needs no further action and does not need to be stored
- It should be retained for later analysis and/or record keeping
- It requires an immediate response
The mission of edge computing is to distinguish between these types of data, identify what level of response is required and act on it accordingly. In most cases it’s far more efficient to perform these functions right there at the edge, where the data is being collected.
When outlier data appears, action may need to be taken. Edge computing can provide a near real-time response to local events thanks to its physical proximity and resulting low latency. No round-trip of data from the edge to the cloud and back again is needed. In addition, the reduced flow of data over the network can produce substantial savings in bandwidth and thus significantly lower networking costs, especially for wireless cellular connections.
Digi Solutions for Edge Compute
Digi delivers a broad product selection for building edge intelligence into applications and connecting edge devices with high-performance connectivity:
System-on-Modules such as the Digi ConnectCore® 8X provide multiple processing units to perform AI and Computer Vision tasks at the edge of the network. Powerful GPU and CPU cores to run neural networks, and support for OpenCL/OpenCV for machine learning and machine vision applications, enable real time performance at low power at the edge. The compact (40 mm x 45 mm) Digi ConnectCore 8X, a system-on-module (SOM) based on the NXP i.MX 8X application processor, also offers the benefits of the Digi SMTplus® surface-mount form factor, which reduces manufacturing costs while increasing design flexibility.
These smart modems have MicroPython integration, enabling embedded developers to fully control the behaviors of their deployed devices' edge compute functionality. Developers today are using the full range of Digi XBee Tools to integrate business logic into Digi XBee 3 modules to dramatically enhance the capabilities and efficiency of their IoT projects.
Whether wired or cellular, these devices perform the edge compute gateway functions of aggregating data, converting it from analog into digital and encrypting it before transmitting it over the network. The volume of data at this point is at its maximum, especially in use cases where hundreds of sensors at scale are gathering data simultaneously. For that reason, the router also filters and compresses the data to minimize bandwidth requirements. Digi routers can also operate offline. For example, if deployed at a remote location, they could collect and store data periodically, then re-establish a connection at regular intervals to transmit information, as needed. A permanent connection to the network is not required.
Digi’s IoT device management platform enables developers to integrate edge functionality into their deployments. With Digi RM, teams can quickly push edge functionality out to their remote devices via firmware updates. Additionally Digi RM integrates with cloud and edge platforms like Amazon Web Services and Microsoft Azure, providing a management interface that integrates the entire compute stack from the data center to the edge.
Edge Computing in Action — Use Cases
The huge expansion of IoT has driven a corresponding expansion in edge computing capabilities and use cases. The following represent just a fraction of the growing spectrum of edge computing applications.
- Manufacturing: Adaptive diagnostics in an industrial setting can be improve the uptime of machines and equipment, cutting service expenses. Edge-compute-generated error codes, combined with historical repair information can provide context for technicians, speeding up troubleshooting and repairs.
- Smart Cities: Edge compute enables public buildings and facilities to be monitored for greater efficiency in lighting, heating and more. In traffic management applications, cameras and signals can improve safety and traffic flow. In the near future, autonomous vehicles, where near-zero latency is critical, will be the most visible and dramatic examples of real-time edge computing.
- Healthcare: Wearable devices can store information on heart rate, temperature, and other metrics, then provide reminders for medication. In addition, edge computing enables developers to ensure sensitive data, such as medical imagery, does leave the device to enhance security and privacy.
Living on the Edge
In 2018, less than 10 percent of enterprise data was created and processed at the edge. Analyst firm Gartner expects that by 2025, that number will reach 75 percent.1 Thus, a lot of organizations that are not using edge compute now, soon will be.
For a successful edge computing solution, it’s important to choose devices that are durable enough to function reliably for extended periods — often years — in harsh edge environments. It’s also important to work with a partner who has both the experience and the expertise to assemble the hardware and software needed to make up such a solution.
Digi can help you with any aspect of your edge compute planning and deployment, from defining a strategy, to programming edge intelligence, to building out your solution. Contact us to start the conversation.
1 Rob van der Meulen, “Edge computing promises near real-time insights and facilitates localized actions,” Smarter with Gartner, October 3, 2018, https://www.gartner.com/smarterwithgartner/what-edge-computing-means-for-infrastructure-and-operations-leaders/