Edge computing is an emerging computing paradigm which refers to a range of networks and devices at or near the user. Edge is about processing data closer to where it’s being generated, enabling processing at greater speeds and volumes, leading to greater action-led results in real time.
It offers some unique advantages over traditional models, where computing power is centralized at an on-premise data center. Putting compute at the edge allows companies to improve how they manage and use physical assets and create new interactive, human experiences. Some examples of edge use cases include self-driving cars, autonomous robots, smart equipment data and automated retail.
Possible components of edge include:
Edge devices: We already use devices that do edge computing every day—like smart speakers, watches and phones – devices which are locally collecting and processing data while touching the physical world. Internet of Things (IoT) devices, point of sales (POS) systems, robots, vehicles and sensors can all be edge devices—if they compute locally and talk to the cloud.
Network edge: Edge computing doesn’t require a separate “edge network” to exist (it could be located on individual edge devices or a router, for example). When a separate network is involved, this is just another location in the continuum between users and the cloud and this is where 5G can come into play. 5G brings extremely powerful wireless connectivity to edge computing with low latency and high cellular speed, which brings exciting opportunities like autonomous drones, remote telesurgery, smart city projects and much more. The network edge can be particularly useful in cases where it is too costly and complicated to put compute on premises and yet high responsiveness is required (meaning the cloud is too distant).
On-premises infrastructure: These are for managing local systems and connecting to the network and could be servers, routers, containers, hubs or bridges.
Edge computing –3 steps for maximizing value
Research reveals four different adoption types, along with their relative successes and challenges, and a three-step framework for maximizing edge value. Read more.
Much of today’s computing already happens at the edge in places like hospitals, factories and retail locations, processing the most sensitive data and powering critical systems that must function reliably and safely. These places require solutions with low latency that do not need a network connection. What makes edge so exciting is the potential it has for transforming business across every industry and function, from customer engagement and marketing to production and back-office operations. In all cases, edge helps make business functions proactive and adaptive—often in real-time—leading to new, optimized experiences for people.
Edge allows businesses to bring the digital world into the physical. Bringing online data and algorithms into brick-and-mortar stores to improve retail experiences. Creating systems that workers can train and situations where workers can learn from machines. Designing smart environments that look out for our safety and comfort. What these examples all have in common is edge computing, which is enabling companies to run applications with the most critical reliability, real-time and data requirements directly on-site. Ultimately, this allows companies to innovate faster, stand up new products and services more quickly and opens up possibilities for the creation of new revenue streams.
What makes edge so exciting is the potential it has for transforming business across every industry and function, from customer engagement and marketing to production and back-office operations.
What makes edge so exciting is the potential it has for transforming business across every industry and function.
Edge unlocks valuable data to shape new opportunities and innovation for the future.
Edge computing combined with other technologies
Edge integrates centralized and distributed architectures. Cloud and the edge work hand in hand to enable new experiences. Data is generated or collected in many locations and then moved to the cloud, where computing is centralized, making it easier and cheaper to process data together in one place and at scale. Edge computing uses locally generated data to enable real-time responsiveness to create new experiences, while at the same time controlling sensitive data and reducing costs of data transmission to the cloud. Edge reduces latency, meaning it lowers response time by doing the work close to the source instead of sending it to the more distant cloud and then waiting for a response.
5G makes edge implementations seamless by guaranteeing the transmission of critical control messages that enable devices to make autonomous decisions. This last-mile technology connects the edge to the internet backhaul and ensures that edge devices have the right software-defined network configurations to do the right things.
IoT and connected devices are unique data sources that need to be secured and registered in the cloud. Edge will reside near or on these data sources.
Containers provide a standardized deployment environment for developers to build and package applications. Containers can be deployed on various hardware, regardless of device capabilities, settings and configurations.
Service and data mesh provide a way to deploy and query data and services distributed across containers and datastores across the edge. These meshes present a single interface that abstracts away the routing and management of services and data interfaces. This critical enabler makes possible bulk queries for entire populations within the edge, rather than on each device.
Software-defined networking allows users to configure the overlay networks. It also makes it easy to customize routing and bandwidth to determine how to connect edge devices to each other to the cloud.
Digital twin is a critical enabler that organizes physical-to-digital and cloud-to-edge. The twin allows data and applications to be configured using domain terms around assets and production lines rather than database tables and message streams. Digital twins allow domain experts (rather than software engineers) to configure applications to sense, think and act on the edge.
Other technologies like AI and blockchain also make edge more powerful. For example, when AI acts on data at the edge, it reduces the need for centralized compute power. Edge also makes blockchain better as more reliable data leads to greater trust and less chance of human error. Data can be captured and relayed directly by machines in real-time, and the increased use of sensors and cameras on the edge means more and richer data will become available to analyze and act on. Edge is also leading a revolution in automation, moving from systematic processes in closed, controlled environments like factories to complex performances in open, uncontrolled environments like agriculture.
The Convergence of 5G, Edge and Cloud
Accenture’s Jennifer McLaughlin and Teresa Tung discuss how 5G, edge and cloud will impact all industries in the coming decade.
Combined with cloud, edge will enable businesses to reimagine experiences. The potential applications of edge have expanded far beyond just manufacturing and IoT. Edge can be incorporated to drive rapid decision-making and improve user experiences by increasing relevance at each touchpoint. Now, edge is helping create new insights and experiences, enabled by the larger cloud backbone.
Some benefits of edge computing include:
Rapid response: Data transmission takes time. In some use cases—like self-driving cars or telesurgery—there isn’t time available to wait for data to make a round trip to the cloud and back. Edge makes sense for these cases where there are requirements for real-time or extremely rapid results.
High data volume: While the cloud can handle very high data volumes, there is a significant cost of transmission and physical limitations of network capacity to take into account. In these cases, it might make more sense to process the data at the edge.
Privacy: Users may prefer (or be required) to keep control of sensitive data locally without sending it to the cloud.
Remote areas: Some use cases are “remote” in the sense of connectivity, whether actually remote (like an offshore oil drilling platform) or practically remote (involving mobile or transportation-related scenarios using edge).
Cost sensitivity: Processing data in different parts of the cloud continuum involves different cost profiles, which can be optimized to minimize total cost across the system as a whole.
Autonomous operations: Where connectivity to the cloud is not possible—or likely to be intermittent or unreliable—users may need end-to-end processing within the local environment to keep operations up and running.
The prime advantage of edge computing is clear: User experience improves because relevance increases with edge. Additionally, edge unlocks valuable data to shape new opportunities and innovation for the future. More sensors generate more data, and there is more processing at the location where the data is created—which is faster, more reliable and safer. Integrated with knowledge from the cloud, the system yields better predictions and more relevant information, repeating in a cycle of continuous improvement.
Other characteristics of edge use cases include:
Intelligent machines and real-time productivity: Edge lets users process data with velocity, enabling robots and sensors to make split-second decisions and complete tasks in smarter, faster and safer ways. This is revolutionizing everything from smart signage to assembly-line quality assurance.
Optimized close to consumption: Digital production and consumption is optimized for the best experience and lowest cost, making edge work for content delivery, for example, or on an offshore oil well.
Experience with extended reality: These use cases can incorporate digital twins and optimize rich experiences in healthcare, the workforce and entertainment, from smart health to mixed-reality gaming.
Privacy and security by default: By processing sensitive data on the edge, these use cases improve reliability and protect privacy. Examples include wearable health devices and the processing of regulated data.
Always-on and untethered: Edge allows for decision-making and processing independent of connectivity for mission-critical and remote applications, like POS or autonomous operations.
Edge computing examples
Let’s dive into a couple of examples of edge use cases that are already happening today and will only improve with a greater 5G rollout and other innovations.
Retail: A flexible, customer-centered experience that is at the heart of a Store of Tomorrow concept, a new integrated vision for the near future of retailing. Edge technology will be a core retail capability in the near future and is a key enabler component for the human-centered experiences at the heart of this model. One of the applications of edge is frictionless store checkout. Long lines are the bane of shops: 86% of consumers have left a store because of them, resulting in an estimated $37.7 billion in missed sales annually in the United States.
An edge network in the store processes data collected by on-site cameras using AI that is trained to recognize inventory items, allowing customers to walk out of the store past a kiosk that accurately charges their accounts without waiting in line. Retailers can provide a superior customer experience, prevent theft and better manage their inventories and supply chains.
Healthcare: Robot-assisted surgery makes the experience easier for surgeons and the procedures less invasive and shorter for patients. Edge computing in this context results in several small changes that add up to having a big impact: The incisions are smaller and the surgeon no longer needs to stand, has a better view of the site and can use controls that are more natural and intuitive.
The prime advantage of edge computing is clear: User experience improves because relevance increases with edge.
Edge computing challenges and opportunities
Organizations looking to realize the benefits of edge computing sometimes face barriers to adoption. Determining the right edge strategy is not easy, but it’s important to experiment—continually refining the approach to set your business on the path toward success. The most common challenges we see are:
Lack of standard and integrated architectures: To get up and running with edge requires the right infrastructure (e.g., cloud provider(s), network, devices). Often, enterprises use multiple, incompatible tech stacks that have to be aligned for edge to work optimally.
Fast-moving ecosystem with multiple tech options: The universe of potential partners and technology is vast, and critical decisions must be made. Continued innovation in network capabilities like MEC and 5G is further complicating the landscape.
Unrealized business value at the edge: It can be difficult for organizations to understand the full business value which can be unlocked by solutions at the edge. Companies must move beyond easy-win use cases that drive quick returns to investments in desirable, feasible and viable experiences for edge computing that delivers sustained ROI.
Innovation fatigue and pilot purgatory: Industrializing and scaling edge solutions for true value can be daunting, and often organizations are too hardwired to quickly flex and scale beyond proof of concept.
Lack of cloud talent to understand what belongs at the edge, why and when: Edge isn’t about retooling, especially for companies that are already leveraging the cloud. It’s about extending those capabilities out to the edge. If you have existing cloud talent, you can leverage their skills to deploy at the edge—the hardware connection is the simple part.
Unique security challenges at the edge: Security has to extend seamlessly from cloud to all possible edge instances, but security in the IoT and edge domain is very different to security in the IT domain. There are many time-critical, safety-critical, and autonomous operations at the edge. Security models take the long design life and legacy infrastructure of devices used at the edge. They quickly become comparatively obsolete and rapid patching may be impossible if production or safety is impacted by reboots. Additionally, devices might be located remotely or in untrusted environments, which requires a blend of cyber and physical defenses. Heterogenous hardware, software and network combinations complicate the rollout of security updates.
Accenture offers a full spectrum of services to help maximize the benefits of edge computing.
Edge computing—or just “edge”— moves computer storage and processing (now often just called “compute”) to the edge of the network. This is where it is closest to users and devices and most critically, as close as possible to data sources.
One of the most cutting-edge applications of edge is frictionless store checkout in retail, allowing customers to pick up items off the shelves and walk out the door, getting checked out without waiting in line.
Cloud accelerates the edge, which enables new experiences. Cloud and edge computing are distinct but complementary. Centrally, cloud brings data together to create new analytics and applications, which can be distributed on the edge — residing on-site or with the customer. That, in turn, generates more data that feeds back into the cloud to optimize the experience.