Your architecture needs an Edge

Introducing the biggest computing paradigm shift in a decade.

Edge Computing has the potential to change the Internet forever. But what is it? Much like Cloud Computing over a decade ago, Edge Computing suffers from competing definitions, misappropriation, and misuse. This article aims to clarify what Edge is, and how it could fundamentally reshape how we build, deploy, and scale software applications.

Where is the Edge?

Edge Computing is bigger than the Internet of Things, and it’s more than a “mini cloud.” The Edge is everywhere. It’s in your car, at the office, and down the street. The Edge can be just about anywhere outside your typical data center. It’s where real-world devices, like phones, cameras, sensors, and drones, meet the network that connects them.

Some experts distinguish between Device Edge and Infrastructure Edge.

Device Edge is simply the end device itself. It’s the phone, the camera, the sensor, or the drone.

Infrastructure Edge, on the other hand, is typically one “network hop” away from the device. It’s the base of the cell tower, the IT closet in your office building, the server rack in the hospital basement, or the WiFi access point. The Infrastructure Edge is often invisible to end-users, yet it is crucial for enabling intensive applications too powerful to run on the Device Edge.

What is Edge Computing?

Edge Computing places compute and storage capacity in these far-flung Edge locations. This idea is nothing new. Placing computers in homes, restaurants, and factories certainly predates the modern data center.

But until now, the Edge Computing landscape has been fragmented and driven by custom, use case-specific solutions. Renewed interest in Edge Computing, spurred by the advent of IoT and Machine Learning, is shifting focus from domain-specific Edge architectures to thinking of Edge Computing as an extension of Cloud Computing.

By treating the Edge like an extension of the Cloud, application developers and operators can loosely retain the elasticity, programmability, and multi-tenancy of the Cloud. End-users can enjoy the benefits of lower latency, cheaper bandwidth, tighter security, and better overall system resiliency and scalability afforded by the Edge.

Edge vs Cloud

Why bother with Edge at all? Why not keep our workloads and data entirely in the Cloud? Simply, not every application belongs in the Cloud.

Edge Computing enthusiasts usually cite use cases that require low-latency feedback loops or that consume massive amounts of backhaul bandwidth (e.g. augmented reality). However, as public cloud providers peer closer and closer to the true Edge, latency and bandwidth concerns lose some of their importance.

System architects will adopt Edge Computing not just because of latency and bandwidth concerns, but because of their requirements for security, resiliency, and scalability that only Edge can provide.

The Future of the Edge

The Cloud to Edge spectrum further compounds the multi-data-center, multi-cloud problem space with a new multi-edge dimension. Designing robust system architectures that can span the globe is an art form, and requires tackling the challenges of scale, resiliency, and security that Edge presents. Like Cloud Computing, companies that learn to embrace Edge Computing now will be well-positioned for a more distributed future. Whether you are a developer, architect, product manager, or CTO, Edge Computing is a trend to know.