QR code separating
Edge Computing
Rapid, private, and secure processing
By Tracy Barbour

loud computing has a silver lining, of sorts, at its edge. Edge computing is an alternative to transferring data to a distributed cloud, which has limitations on bandwidth (transmission capacity) and latency (transmission delay). As networked devices become more numerous and powerful, edge computing is steadily growing in usage. By 2025, 75 percent of enterprise data will be processed at the edge, compared to only 10 percent today, according to technology research firm Gartner, Inc.

Edge computing is related to another somewhat recent technology, the Internet of Things (IoT), according to Kenrick Mock, a professor of computer science and dean of the UAA College of Engineering. “The central premise behind IoT is to have everyday objects and sensors connected to the internet,” he says. “Under the vision of IoT, your watch, thermostat, garage door, oven, refrigerator, and even your coffee mug could all connect, communicate, and compute via the Internet. In edge computing, there is the same vision of many interconnected computing devices, but the distinction is where the computation occurs. If you think of large, powerful, remote servers as being in the ‘center’ of the cloud that makes up the Internet, then the ‘edge’ of the Internet are devices on the periphery, such as your phone, laptop, thermostat, watch, or sensor.”

The edge is not a specific location; it’s about facilitating the distribution of services to where people need them. Regardless of the device, the edge is located near the user. “Edge computing puts that processing, the brains, or content as close to the users as possible,” says Victor Esposito, vice president of engineering and architecture at GCI. “It’s as close to an instantaneous response or transaction as you can get.”

Since edge computing brings processing capabilities nearer to the user, it eliminates the trip to the cloud data center. That significantly reduces latency, the amount of time it takes for data to pass from one point to another. Edge computing allows for quicker and more comprehensive data analysis and improved customer experiences. Decreasing transaction time is especially important in an isolated place like Alaska, where communication throughout the state and the rest of the world involves unavoidable transmission delays to the Lower 48. “For us, bringing content into Alaska and getting it closer to our customers is important to giving people that traditional urban experience,” Esposito says.

Kenrick Mock
UAA College of Engineering
Common Applications

Edge computing has been around long enough to be employed in an array of products and services. Some common examples are smart devices that run code locally instead of in the cloud; medical monitoring devices where real-time response is critical; self-driving cars that need to make split-second decisions; and interactive video conferencing, gaming, or streaming that requires a significant amount of bandwidth.

Victor Esposito

In Alaska, edge computing is being deployed in business scenarios at enterprise scale. It’s primarily used today with larger firms that require the lower latency that edge computing provides, Mock says. Alaska Airlines, for instance, uses multi-access edge computing (MEC)—edge computing on the telecommunications network—to quickly access airline, baggage, and ground operations information. Consumer electronics are also moving in this direction. “For example, Amazon’s newer Echo devices will now analyze audio on the device itself, whereas older devices required audio to be transmitted to a remote server for processing,” he says. “Local analysis allows a faster response and the ability to follow up with a more natural dialogue.”

The healthcare and education industries in Alaska are also leveraging edge computing within. Lower latency allows clinics to connect directly with patients and schools to connect directly to students. As a service provider, GCI ensures its clients have the low-latency connectivity to support their efforts in these areas, Esposito says. “For our healthcare clinics and education businesses, we make sure we provide them high bandwidth and highly reliable connectivity from their remote locations to the main hospitals and to their students,” he explains.

Evolution of Edge Computing

Like any technology, edge computing is perpetually evolving. In the early days of computing, centralized applications ran on isolated, bulky mainframe computers. Then personal computing allowed for decentralized applications that run locally on the user’s device. With more recent cloud computing, centralized applications run in cloud-based data centers that can be accessed from any device over the internet. Edge computing closes the distance, running centralized applications close to users on the device itself or on the network edge.

“I anticipate that in the next few years, you’ll start to see edge compute as a service. As a business, you may buy it, so you have a whole package—connectivity, high bandwidth, low latency, and cloud services. Bringing these ideas together, you’ve got the compute right on the edge of your links.”
Victor Esposito
Vice President of Engineering and Architecture

Esposito says computing has, in a way, come full circle, making a virtual connection back to the mainframe. “Initially we placed compute toward the edge out of necessity,” he explains. “Then we had a massive step forward with the internet, so people were able to pull that compute into centralized locations (data centers). Because of virtualization and high-bandwidth links, we now have choices. We can put compute out toward the edge, virtualize it, and use it for many applications and functions, or move applications into the cloud.”

Mock sees Alaska organizations employing edge computing at a similar rate as the rest of the country as part of normal technology upgrades. “Our networking and telecommunications firms are developing capacity in this space so that edge computing applications can be realized, as the computing is often needed at a base station or local data center,” he says. “As computation is pushed to the edge, we also need more powerful and capable devices.”

He adds: “Integrating edge computing with 5G networks and traditional computer networks is also occurring, especially as expectations rise to seamlessly access services from a phone or computer. There are a host of AI [artificial intelligence] applications that are enabled by edge computing. For example, security cameras with built-in image processing can detect if detected motion is a potential human intruder or a wandering moose. For resiliency, it is better if this analysis can occur locally rather than be dependent on a connection to a remote internet server.”

The roll out of 5G communications networks, with WiFi speed at cell phone range, is accelerating the usefulness of edge computing. According to Esposito, 5G is mostly intended to be the framework for virtualized edge deployment. “The network function is being virtualized [simulated in a computing environment] and pushed to the edge,” he explains. “The great thing about virtualization and compute is they work together. When we have a failure at the edge, that can be backed up in a data center in the Lower 48.”

Esposito expects developments in areas like IoT, AI, and 5G to continue advancing with edge computing. “It’s really making sure the compute is right on the other side of that low-latency link so the AI or IoT can respond quickly and react,” he says. “I anticipate that in the next few years, you’ll start to see edge compute as a service. As a business, you may buy it, so you have a whole package—connectivity, high bandwidth, low latency, and cloud services. Bringing these ideas together, you’ve got the compute right on the edge of your links. That’s where we see a lot of possibility in the future.”

Mock also anticipates growth of edge computing. The proliferation of IoT devices continues, and edge computing is likely to grow with it, he says. There will be more wearable computing—particularly medical devices—and smart appliances around the home. “AI has made great progress, especially around image recognition but also with understanding human language, so expect more natural interaction with computing devices,” he says. “We are also at the point where edge computing can enable the vision of a smart city. Examples include real-time monitoring of traffic and potholes, so your car knows the best route to drive; streetlights that turn on when needed rather than on a timer; or the ability for your car to direct you to an empty parking spot rather than driving around the block looking for one.”

Edge computing will be mostly “invisible” to consumers, Mock says, but they will benefit from newer and faster services. “As more devices are added to the IoT, we will need a commensurate rise in edge computing to avoid bottlenecks at centralized servers,” he says.

Expanding Edge Solutions

Technology companies like AT&T and IBM are pivoting to create edge solutions that enterprises can leverage easily. AT&T, which has been deploying private cellular networks for businesses, universities, and the public sector for years, is working on bringing private 4G/5G wireless networks as an integrated platform with connectivity and applications to enable low-latency services at the edge. Its upcoming service, called AT&T Private 5G Edge, allows users to roam beyond the geographic boundaries of the AT&T private network while staying connected through the AT&T public network, according to a February 24 press release.

“As more devices are added to the [Internet of Things], we will need a commensurate rise in edge computing to avoid bottlenecks at centralized servers.”
Kenrick Mock
UAA College of Engineering

AT&T Private 5G Edge, which is currently under development with Microsoft, uses Azure private MEC to deploy private wireless networks rapidly across radio spectrums, including Citizens Broadband Radio Service (CBRS). The service is ideal for companies and organizations where private networks need to be simple, flexible, and easy to use. For example, a hospital might use its private network to closely track ventilators, wheelchairs, and other critical items in its building. But if a ventilator is on loan to another hospital, the company’s roaming capability could ensure that machine always remains accounted for even outside the private network.

“With AT&T Private 5G Edge, we are enabling customers to create and deliver innovation faster—with simplicity, flexibility, security, and high-speed wireless connectivity,” Rupesh Chokshi, vice president product strategy and innovation at AT&T Business said in the press release. “This solution opens the door to entirely new applications and use cases we haven’t even imagined yet.”

In an interview, Chokshi says he thinks there is huge demand for AT&T Private 5G Edge in rural areas. “Oil drillers, for example, in extremely remote regions could use a private network running on CBRS to enable on-premises, low-latency edge services they otherwise wouldn’t have access to,” Chokshi says. “That could either be a self-contained CBRS-based private network or potentially have connectivity back to the public network either via a roaming agreement with another carrier or through AT&T’s commercial network if it’s available in that area. Overall, we think AT&T Private 5G Edge will be appealing to midsize companies with small facilities or to larger companies that have a large, distributed footprint (such as a retailer with locations across a region or across the country), or a university with multiple campuses.”

Last year, IBM released several solutions to help businesses take advantage of 5G and edge computing. They include IBM Edge Application Manager, an autonomous management solution to enable AI, analytics, and IoT enterprise workloads to be deployed and remotely managed; IBM Telco Network Cloud Manager, which facilitates intelligent automation capabilities to orchestrate virtual and container network functions in minutes; and a portfolio of edge-enabled applications and services. These services, according to IBM’s website, are all designed to allow companies across industries to realize the benefits of edge computing, including running AI and analytics at the edge to achieve insights closer to where the work is done.

Choski Rupesh
AT&T Business

In Alaska, GCI is also launching technology to support edge-enabled devices. Recently, GCI became one of the first companies to deploy Remote MACPHY Device (RMD) in its broadband network. RMD service—in partnership with CommScope—is a component of GCI’s broadband internet. According to Esposito, RMD technology allows GCI to move internet distribution from the centralized headend out to neighborhood nodes, enabling the company to provide more reliable, better quality, and higher performance service. “Between that and our fiber-to-the-home equipment or our 5G networks, we’re employing the latest technology for people to use to provide that low-latency, high-bandwidth platform,” he says.

“From edge compute for gaming or video content, we bring that content into Alaska and push that into our data centers, so it’s close to our subscribers,” Esposito says. “As a provider across Alaska, we’ve got an obligation to make sure people have the tools and connectivity they need to run their business.”