By Art King, Corning
Edge computing: What is it? At the end of this article, we hope you’ll be able to answer that question. This was written to provide an understanding of the concept, how it underpins 5G, and how it will influence both mobile operators and enterprises.
Today, enterprises and internet providers use edge computing as a building block to improve performance. Before we explore 5G and edge computing, let’s look at a few of the existing examples to help us understand why it is used.
· Enterprise email systems deploy edge computers to insulate their employees from the wait time of uploading/downloading large files from a distant mail server as the perceived slowness of the system can trigger performance complaints. If the “edge” of the email system is local, any large files are immediately sent-to or received-from the email system and this hides the potentially long wait times to transmit or receive large files across the network.
· Internet content distribution networks (CDN) have edge computing nodes installed across the world in wired and cellular networks. Web properties arrange to store digital objects in the distributed nodes as “CDN enabled content.” When they are requested, the CDN determines the nearest edge node to the requestor and serves the content to them from that edge node. The CDN, by delivering content from the edge, assures that the web page loads as fast as possible, no matter where the user is located.
From the examples above, it is clear that edge computing reduces the latency between a device and service where latency is the roundtrip time between two systems over a network. While reducing latency is one of the its main benefits, edge computing has other benefits that will be exposed as 5G use cases are discussed.
Edge computing and 5G
We are on the cusp of introducing computing elements into the cellular radio access network (RAN). As part of the 5G specifications, the Ultra Reliable Low Latency Communications (URLLC) building block includes edge computing as a necessary enabler. While earlier initiatives had introduced edge computing in the LTE RAN, there was a lack of specifications and industry buy-in on the concept that prevented large-scale commercial deployment.
Edge computing pilots started in 2011 and, with 5G, it has been architected into the cellular network. In 2014, the European Telecommunications Standards Institute (ETSI) created the multi-access edge computing (MEC) working group, with 85 member companies, to create an application server framework for bringing applications to the edge of the network. MEC includes methods for applications to incorporate radio information, collect usage metrics, and other common operations functions. Much of this work is being carried forward into 5G. In the 5G domain, industry and vertical market players are making significant investments in edge computing to put it to work. We’ll frame the impact of these trends and then diagram the technology.
As wireless is embedded into more and more systems, the sheer volume of traffic that will be generated is sure to stress centralized systems and trigger the decomposition of them such that computing instances associated with the main system will be deployed at the edges of networks. The goal is for the edge to do as much work as possible locally before the cloud service has to be contacted.
This decomposition to edge computing is already in development as the major enterprise cloud service vendors have all introduced edge platforms and interfaces to enable their enterprise customers to decompose their cloud applications and distribute the application edge. While this makes great sense for an enterprise global network for performance management, moving the strategy into cellular infrastructure takes more planning.
A number of factors come into play with edge computing.
· Mobile devices finding and connecting to nearest edge application instance
· Compatibility of the edge computing container to host the edge application
· Where to locate the edge computing systems for best cost/performance
From a practical perspective, the enterprise cloud service vendors’ edge computing environments have a number of location choices available to them.
Edge location choices:
· Easy: Extend multi-enterprise cloud edge systems to the mobile operator’s core data centers. Any installed apps on mobile devices can utilize the cloud edge in the data center instead of going to the cloud. In this instance, just a few locations in the mobile network’s data centers serve all the mobile devices in that network.
· Well planned: Extend multi-enterprise cloud edge systems adjacent to the RAN in areas where enterprises have higher device density or more traffic like urban areas or known headquarters locations. This strategy also includes the same mobile network data centers from the prior option. This choice helps improve mobile device app performance from the RAN collocated edge system or, should it not exist, the cloud edge in the mobile network’s data centers.
· Ubiquitous: Extend multi-enterprise cloud edge systems adjacent to the RAN in all areas nationally where device density exceeds a targeted threshold. Like the “well planned” choice, this strategy includes the same mobile network data centers from the first option. This choice helps improve mobile device app performance in urban/suburban areas or, should density be below threshold, the cloud edge in the mobile network’s data centers.
If we assume multi-enterprise edge computing is relatively easy for enterprise app builders to embrace, while also providing visible improvements in app performance and engagement metrics, the gradual rollout to each tier of the major mobile networks seems justified. This combines with a commercially viable cost model in which the enterprise cloud service vendors bill enterprises for inclusion of edge systems into mobile apps.
Other industrial clouds with specialized goals may also develop that will leverage the 5G edge. For example, self-driving vehicles and healthcare will most likely converge on a common set of standards and infrastructure to drive economies of scale and widespread industry collaboration so that, like the underlying 5G providers, these industries can be enjoyed across the globe and not be limited to national markets.
What about indoors?
The assumption of 100% mobile device ownership has transformed our design thinking for enterprise mobility. In the drive to increase access to amenities in workspaces, there is an opportunity to shape a strategy around mobile devices. We see this emerging now in co-working spaces and enterprise remodels that drive for the same vibe as co-working. In these environments, an app can check you in, unlock doors, book conference rooms, request resources, and help with wayfinding (getting from present location to destination). In the near future, this capability can be a service offered by the facility to all tenants or across an enterprise.
Additional, our mobile devices now enable text-aware computing. The Gartner IT Glossary defines context-aware computing as “a style of computing in which situational and environmental information about people, places and things is used to anticipate immediate needs and proactively offer enriched, situation-aware and usable content, functions and experiences.”
If you assume that every mobile device acts as a digital proxy for the owner, this is a driver for implementing 5G edge computing to enhance the experience in the building and potentially offer local APIs to enterprise business applications so they can add a layer of context awareness.
Context-aware decisions primarily revolve around two questions: 1) Are you a tenant or a public subscriber passing through a building? 2) Are you present in the building?
Many potential context-aware applications will use simple “presence,” meaning the mobile is present on the building’s RAN. Enterprise applications will be able to consume information about whether a device is within the facility, where it is located, and act accordingly. We foresee that enterprise directories will use AI with automated provisioning to ease the task of identifying employee mobiles. For instance, if the 5G edge system sees the arrival and departure of the same device every day and it has the enterprise’s mobility management onboard, it can be assumed that it’s an employee based in the building.
Using the 5G edge capability
Energy management is one example where location information can be leveraged by the enterprise building management to optimize resource consumption by reducing dependency on simple timers or thermostats controlling lighting, HVAC, and other services. With building management taking actions as devices appear within the building and their associated density utility consumption can be managed along with increasing comfort. Device density information can be used to optimize heating and cooling based on amount of people per square foot in a zone. Imagine not having to adjust a conference room thermostat, because it automatically adjusts to maintain optimal room temperature based on detected occupancy. This simple function provides ROI by managing both comfort and utility use.
For public safety protocol in large commercial and government buildings, location information can enhance process capabilities in emergencies. There are a number of useful capabilities that would be valuable to tenants and first responders.
Here are some brief examples. In an emergency, it would be possible to text every device that is present and, by knowing their location, edge systems could calculate an ideal route out of the facility. For emergency calls, knowing the location of the caller can get first responders to them faster and, as health wearables improve, if a person had an event that resulted in unconsciousness their mobile device could contact emergency personnel on their behalf. In a fire or other emergencies, the presence and location of devices could enhance rescue efforts. Additionally, incident commanders could reach out to contact people in the building to increase their situational awareness and advise them.
Finally, there are a vast amount of Internet of Things (IoT) services on the horizon that will also leverage context and presence to improve their ability to serve us.
What does edge computing look like in the future?
The diagram shows how edge computing will provide the local breakout necessary to both route data locally and provide a presence event stream for building management and enterprise tenant systems to subscribe to. This illustrates one possible future for system architects and developers to think about.
Many of the functions are part of decomposed cloud services and shared across many enterprise buildings so that the services are universally available.
What can we do today to prepare?
Build-out cellular infrastructure. It’s an uneven world with some buildings having service improvements and many with nothing. Pressure continues to mount for robust indoor LTE coverage in poorly covered buildings by both current and prospective tenants.
And, not to worry, indoor cellular improvements for LTE are an investment in the future because they will carry over to 5G. Why? Well, 5G in the United States will be rolled out in a mode called 5G-NSA where NSA means “non-standalone.” In other words, 5G-NSA under the hood is LTE combined with 5G radio technology. An investment in LTE today is protected as it will be upgraded to 5G via the addition of 5G radios and edge computing in the future.
Finally, developing a multi-year roadmap to 5G with contingency plans is necessary to reach the desired destination. Many enterprises have access to resources at the mobile operators and key infrastructure software providers to help develop their roadmaps and clarify the business cases. This will be a journey for the industry, but we are seeing that the mobile operators, cloud providers, and software firms targeting this market will act as the “glue” amongst the disparate systems and edge compute platforms to help realize the vision.
For those of us who pioneered edge applications early on, it is encouraging to see edge computing architected into 5G. The future ability to loosely connect mobility to both building management and enterprise systems will positively transform the employee experience and increase the sustainability of the building itself. Good times lie ahead.
Art King is director of enterprise services and technologies with Corning Inc.