​​Edge Devices, Edge Computing, and Edge IoT Explained

Edge devices are derived from the notion of a "network edge" — that is, the terminal or end point of a network's reach. Edge devices bridge the gap between the network edge and the external world. Traditionally, that's referred to network infrastructure like edge routers, bridges, and servers. Modernly, edge devices comprise a huge number of computers, network components, servers, and IoT devices. The definition of edge device has expanded to accommodate devices deployed across a tremendous variety of evolving industries and use cases.

What are Edge Devices?

Typically, an edge device is a computer (e.g., a server), piece of network infrastructure, or IoT device that exists at the edge of a network, acting as an interface between the external world and that network. That's a very broad definition, and that's why, usually, the edge is defined from the perspective of a particular entity, be that an internet service provider (ISP), a platform (like a social media service), or an enterprise or other large organization.

Enterprises often use "edge device" to refer to the distributed systems they own that interface between the "outside" world (i.e., the broader internet) and their own internal systems. The edge is a useful concept applied in this way, because it allows an organization to define, for example, the security, performance, and other operational factors that must be achieved "at the edge." 

Edge devices might be a network router, a bridge or gateway, an edge server, or even IoT devices like industrial sensor networks, highly automated retail point-of-sale (POS) systems, or camera-equipped devices running computer vision edge AI applications. Edge IoT devices are the most diverse segment of edge devices, and they comprise the growing overlap between the definition at the "edges" of edge and IoT because of their use of edge computing. But before understanding edge IoT, we need to establish a definition for edge computing.

What is Edge Computing? 

Edge computing refers to computing conducted locally "at the edge," as opposed to sending data to a central computing facility or the cloud. Implied in this definition is that placing computing resources at the edge results in some kind of operational or economic advantage. The rise of edge computing has been driven by a number of factors, including mobile devices at the edge, IoT, and automation across many industries (transportation, logistics, retail, manufacturing, warehousing, and more). Smartphones are an example of one form factor driving edge computing because they are increasingly good at performing complex tasks locally instead of solely relying on cloud services. This aligns closely with the shift toward local processing to conserve bandwidth and reduce latency. This could include conversational AI with speech recognition, object recognition via photo or video, or encoding high-quality video to a format compatible with social media, though it’s worth noting that some truly advanced computing tasks will still require the use of the cloud.

Imagine you run a large social video platform, like TikTok or Instagram. Your users upload a tremendous amount of video data to your platform every minute, and that video likely is delivered in a variety of formats and resolutions. Your platform, however, processes all of those videos into a common format that is optimized for mass distribution (i.e., it is very space efficient and universally supported). All of the video you receive could be processed centrally, but in particularly high-demand regions, you may want to deploy edge computing. Using purpose-built computers designed to process video, you could have users' videos encoded and ready for publication on your platform much more quickly, improving both the user experience and freeing up your own computing infrastructure.

What is Edge IoT?

Edge IoT refers to the edge Internet of Things. IoT encompasses a massive superset of devices — more animal kingdom than genus or family — but IoT at the edge tends to refer to those IoT devices that are responsible for a significant amount of on-device computing. Put another way: Edge IoT devices are IoT devices that engage in edge computing-like use cases.

Imagine you work in IT for a national chain of supermarkets, and you operate a large fleet of self-checkout POS (point of sale) systems. You have an internal network and may even operate servers inside your own infrastructure to deliver various data and services to users and systems across the country. Those self-checkout systems, though, live at the edge of your infrastructure — they are the interface between external users (customers) and your internal infrastructure, locally facilitating checkout operations, recording data for inventory management and accounting, and maybe even capturing video for your loss prevention team. They then send this data from the edge directly to internal systems (likely via a VPN tunnel). This makes them, under the modern definition, edge IoT devices — they are the critical, internet-connected bridge between your locally distributed operations and your internal systems.

The industrial sector also can give us a strong example of IoT at the edge. Imagine you operate a factory producing mechanical components for vehicles. You must produce those parts at a given quality, speed, and consistency to meet your operational goals. Automation of those controls (quality, speed, consistency) is increasingly common, as sensors such as cameras utilize local computer vision algorithms to determine if the output of a given manufacturing system is "in spec." While that system is doing all of the computing necessary to make that determination locally, it also acts as a crucial bridge to the internal infrastructure that can use this information for other operational objectives (e.g., reporting and analytics). This makes it an IoT device that is engaged in edge computing and, thus, an edge IoT device.

The line between edge device and IoT device is continually blurring, and the terms can be used interchangeably to describe certain devices.

What are Edge Servers?

Edge servers deliver services at the network edge. For example, Netflix may build a large edge server facility in a particular region, allowing it to deliver video content to customers in that region by a far more direct network path — increasing performance, and also reducing the overall burden on larger internet infrastructure. An ISP will then engage in an agreement to build out its edge network capacity to supply the throughput (bandwidth) to deliver Netflix's service to customers. This is in the interest of both the ISP — whose large network is unburdened by this more direct delivery route — and the service provider (Netflix) — who delivers higher quality service at an increased capacity. From the perspective of an ISP, every piece of network equipment that connects to the internet is an "edge" device. Because of this, an entity like an ISP doesn't tend to think of the edge in terms of devices, but rather as a demand load to manage — arguably more akin to electricity than bits. ISPs have dedicated edge infrastructure designed to support the needs of customers at the edge, be it streaming video, uploading files, or accessing services in the cloud. ISPs will often engage in infrastructure agreements with service providers, like Netflix, to build out "capacity at the edge" for extremely popular services.

Are Smartphones, Smartwatches, and Laptops Considered Edge Devices?

This is debatable. A smartphone may well engage in edge computing — transcribing speech data to text locally, for example — as might a smartwatch running algorithms to detect activity levels, or a powerful laptop running a large language model. But this is where the definition of "edge device" starts to lose its utility. When effectively every computer can be an edge device depending on the way it is used in a given moment, the term doesn't really serve much of a functional purpose. That said, this increasing blurring of the lines is illustrative of the rapid growth of heterogeneous computing and the high degree of cloud connectivity in our modern world. When your smartphone is collecting data about the world around you to make both locally computed determinations — for example, where you are and how fast you're moving — and also sending that data to services in the cloud for further computation (how many people are also there with you, when will you arrive at your destination), you're really getting into a question that's more philosophical than technical.

Generally, though, most definitions of edge devices don't include these "general purpose" computers we've described above. If a device is not used in a dedicated edge computing capacity but rather does edge computing on a contextual basis, it's probably fair to say that it is not a "true" edge device. Much in the way your car is not a taxi, but potentially could be used as one. With the blurring of these definitions, at least one thing is becoming clearer: The need for more precise terminology.

If You Need to Manage IoT Edge Devices, Esper is for You

Whether you need to manage IoT devices at the edge, a fleet of edge computing devices, or a mixture of the two, Esper has all the tools you need to effectively manage all your hardware. Book a demo with us today to find out how.