The use of edge computing in the enterprise is set to dramatically expand as companies and consumers alike connect more devices to the internet, as superfast 5G network services expand their reach and as organizations pursue opportunities enabled by the technology.
According to a 2019 report from Gartner, more than 50% of large enterprises will deploy at least one edge computing use case to support IoT or immersive experiences by the end of 2021, up from less than 5% in 2019.
The number of edge computing use cases will jump even further in the upcoming years, with Gartner predicting that more than half of large enterprises will have at least six edge computing use cases deployed by the end of 2023. In 2019, only 1% of large organizations had six or more edge computing deployments.
The rise in edge computing rests with the growing availability of technology to handle real-time analysis of data generated by endpoint devices -- analysis that, when combined with AI, machine learning and automation, can be used to control endpoints' actions with limited or no human intervention.
Fully autonomous assembly lines, for example, can detect and then correct mistakes or security checkpoints that can admit authorized users based on their biometrics. Such scenarios need edge computing for its low latency and reliability.
"The rapid development and adaptation of IoT, sensors, mobile devices and other connected devices mean more data and widespread needs for [edge] computing," said Bruce Guptill, chief strategist at Addressable Markets and member of The Analyst Syndicate, a community of senior independent analysts.
"But, with the volumes of data in use, the vast range of data types and formats, the increasing need to protect more types of data in more ways and the need to utilize the data on so many devices of so many types, it's become inefficient to go back and forth to multiple clouds," he said. "And network bandwidth is not keeping up with the trend, so response times get worse."
5 benefits of edge computing
True to its name, edge computing takes compute out of an enterprise's core data center and places it close to endpoint devices where data is being generated, which brings several key benefits, including the following.
Delivering edge computing with purpose-built devices -- a nearby server or virtualized data center next to the endpoints -- eliminates the need to move data from endpoints to the cloud and back again. Decreasing that travel time shaves time off the process, which can be measured in seconds, sometimes even milliseconds. "Since data doesn't have to travel all the way back to a core site -- i.e., the data center or a public cloud provider -- edge applications and services can benefit from real-time or near-real-time levels of latency," said David Williams, managing principal at Ahead, a provider of enterprise cloud solutions.
According to experts, edge computing enables increased security and resiliency because its decentralized nature eliminates a single central point of failure, said Yannis Kalfoglou, AI and blockchain leader at PA Consulting.
As a result, security teams can isolate hacked endpoints and edge computing devices. Moreover, edge devices can have device-specific protocols for security, making it more difficult for bad actors to learn how to hack into the numerous devices.
One of the key benefits of edge computing in the enterprise is increased cost savings, according to Williams. "Edge computing architecture and technologies can often be implemented at a lower cost than their centralized equivalents," he said. "Further savings can also come from a reduction in the connectivity costs as a result of lower amounts of data being sent to and from the edge and the core/central site."
According to experts, edge computing can continue even when communication channels are slow, intermittently available or temporarily down. For example, an energy company with edge computing deployments on an oil rig doesn't have to constantly rely on an available satellite connection to relay data back to a data center for processing -- opting instead to move only the necessary processed information from the edge back to its data center when the connection is available, said Teresa Tung, managing director at Accenture Labs.
If there are any failures at the edge, the impact might only be limited to the affected devices -- the operations overall will still be able to continue, thereby improving the reliability of the entire system.
Like cloud computing, organizations can add edge devices only as they expand their uses so that they're deploying and managing only what they need, said Dan Miklovic, founder and principal analyst at Lean Manufacturing Research LLC and member of The Analyst Syndicate.
Others also noted how the decentralized approach of edge computing makes large-scale deployments more manageable. "It's easier to address the scale of each edge location individually and with a decentralized approach than to address the aggregate at a centralized processing location, [which is] what we call the 'core,'" Williams said. "This is how organizations are supporting thousands, if not tens or hundreds of thousands of endpoints -- numbers [that were] incomprehensible in the centralized model," he said.
7 use cases for edge computing
Each organization has its own combination of factors and motivations for deploying edge computing for particular use cases -- for example, it might require low latency and speed in one case and reliability in another.
Technology leaders and researchers say many organizations across nearly all industries are now deploying or testing use cases for edge computing. Notable use cases include the following.
1. Autonomous vehicles
Autonomous vehicles are a prime edge computing use case, as they can only operate safely and reliably when they're able to analyze all the data required to drive in real time. The volume of data that these vehicles amass is staggering. Industry experts estimated that the data generated by a single autonomous car could be between 5 TB and 20 TB a day. And, while 5G will certainly be able to handle more capacity, the existing 4G network is nowhere near capable of handling all that data at sufficient speeds.
"Autonomous vehicles must aggregate and process huge amounts of data of different types in different ways from multiple sources -- including other vehicles -- [and] more or less instantaneously while in motion," Guptill said.
That necessitates onboard computing power and edge data centers for mission-critical processing for navigation, vehicle-to-vehicle communications and integration with emerging smart cities.
Edge computing will also help civic authorities, such as traffic agencies, public transformation departments and private transportation companies, better manage their vehicle fleets and overall traffic flow by enabling rapid adjustments based on real-time, on-the-ground conditions. For example, edge computing platforms deployed to handle vehicle data can determine which areas are experiencing congestion and then reroute vehicles to lighten traffic.
2. Stronger security
Organizations can use edge computing to enable video monitoring and biometric scanning, as well as other surveillance and authorization measures, where analyzing data in real time is critical to confirming that only authorized individuals and approved activities are taking place. For example, companies can use a biometric security product with optical technologies to perform iris scans with edge devices analyzing those images in real time to confirm matches of workers with authorized access.
Healthcare data is coming from a variety of devices, including those in doctor's offices, in hospitals and on patients themselves. Moving that data to a central location for analysis can create bandwidth congestion -- but all the data doesn't necessarily need to be moved and stored in centralized servers. Every single normal heart rate reading from a patient's medical device, for example, might not need to be retained. However, some pieces of data are so critical that the need to analyze and understand them can't be subject to any delay due to low latency or unreliable network connectivity.
Edge computing can take and process data coming from endpoint medical devices in real time and identify what data points aren't critical -- i.e., the normal heart rate readings -- yet also identify, process and react to the critical data points, thereby alerting clinicians to act on those as quickly as possible.
4. Manufacturing and industrial processes
Industrial IoT has added millions of connected devices in manufacturing plants and other such industries that gather data on production lines, equipment performance and finished products. However, all the data doesn't need to be handled in centralized servers -- every temperature reading from every connected thermometer isn't important.
Rather, most organizations only need to bring aggregate data or average readings back to their central systems, or they only need to know when such readings indicate a problem, such as a temperature on a unit that's out of normal range.
That's what edge computing enables; it gives organizations the ability to take and understand the data so fast that problems can be identified and solved quickly, said Gerald Kleyn, vice president and general manager at Hewlett Packard Enterprise.
He noted that speed is particularly important in manufacturing and similar settings where automated assembly lines move rapidly and require real-time interventions to address problems. For example, he cited one manufacturing plant where edge computing took a second to analyze product quality -- a full 20 seconds faster than the same analysis took when the manufacturing data was moved into the cloud for processing.
5. Augmented reality
Guiding workers through their jobs, training employees on new processes and teaching students complex concepts will increasingly be delivered through headsets that provide a virtual reality or augmented reality learning experience, Miklovic said.
Such experiences could be delivered via centralized computing resources, but costs and latency can lessen UX, whereas edge computing can deliver reliable real-time access to required information at lower costs.
6. Enhanced workplace safety
Advancements in sensors, computer vision and AI are further expanding workplace safety applications, as running these technologies at the edge enables organizations to monitor conditions and identify and be alert to dangerous conditions in real time.
For instance, companies can use locational data from on-site employees to enforce the new social distancing requirements during the COVID-19 pandemic, alerting them if they move and stay too close together. Because such locational data has no value beyond that moment, the information can be collected and processed on the edge rather than moved and stored in the corporate data center.
7. Streaming services
"Over-the-top [OTT] streaming platforms are quickly becoming the standard means for distributing content," Williams said. "While IPTV [IP television] was the original target with content produced and centrally distributed to consuming devices, we see OTT evolving to include original content, live events and even regional content with higher demands for flawless [UX]."
This is a driving factor behind media companies utilizing edge computing capabilities, according to Williams -- it's enabling organizations to reduce the latency, while ensuring high-quality video and streaming performance.
More edge computing opportunities on the horizon
Although such uses of edge computing are already delivering value, experts predict organizations will continue to expand how they implement edge computing to improve current operations and activities, as well as to develop and support new products and services.
"As organizations continue to find ways to harness data collected at the source, they will also continue to provide applications and services that process and consume that data locally," Williams said.
"This 'data gravity' will fuel a new generation of data-driven solutions at the edge that were not previously possible," he continued. "Service providers will continue to make investments that solve the connectivity challenges of today; AI will continue to advance and become more distributed and decentralized between edge and cloud aspects of its technical architecture; [and] we'll also see an accelerated move to open, secure and cloud-native standards with an emphasis on operationalizing edge technologies, from self-resilient to self-healing. The future is looking good at the edge."