The Open Glossary of Edge Computing defines Edge Computing as:
The delivery of computing capabilities to the logical extremes of a network in order to improve the performance, operating cost and reliability of applications and services. By shortenting the distance between devices and the cloud resources serve them, and also reducing network hops, edge computing mitigates the latency and bandwidth constraints of today’s Internet, ushering in new classes of applications. In practical terms, this means distributing new resources and software stacks along the path between today’s centralized data centers and the increasingly large number of devices in the field, concentrated, in particular, but not exclusively, in close proximity to the last mile network, on both the infrastructure and device sides.
The interesting part of this definition is distributing new resources and software stacks along the path… which implies that there a several layers between the logical endpoints of a system and the cloud or data center. And typically as you get closer logical enpoints there are less computing resources available. The following diagram shows the layers that often appear in modern IoT like systems.
Each layer plays a role in solving the problem being addressed by the system. And in some systems not all layers will be present. For example, with many consumer devices, they connect directly to the cloud or centralized data center. See the brief descriptions for each layer later in the article.
Edge computing has had a significant impact on modern computer vision systems. More and more computer systems are being deployed where timely business-critical decisions are being made based on what a camera sees. Often the time it takes to send images or video to a centralized location will not meet the timeliness required to make the necessary decisions. And sometimes the physical limitations of networks don’t provide enough bandwidth to send all of the images or video to a centralized location. The final consideration relates to privacy where it may violate privacy laws by sending images or video to a centralized location.
With edge computing, we are able to bring the processing of the images or video much closer to their source. This will address latency, bandwidth and privacy concerns. And by doing so increases the art of what’s possible with computer vision.
Cloud / Centralized Data Center
One or more physical structures that house large compute, storage and networking resources. These resources are often accessible on-demand and have multiple tenants accessing them. Often located at a significant geographical distance from the endpoints in the system.
Edge Cloud / Local Data Center
A cloud or data center that is capable of being deployed closer to the endpoints or even in the same geographical location as the endpoint devices. Capable of performing the same functions as the cloud or centralized data centers but at a reduced scale based on its compute, storage and networking resources.
Device Edge Cloud
An extension to the cloud or local data center that is capable of being allocated specific workloads but often doesn’t provide elastic resource allocation. For example, an edge server that acts as a historian for all the data being produced by the endpoint devices within the same factory.
Edge computing capabilities typically in the last mile network. This often includes gateway or industrial PCs for compute and storage. And when available will use spare resources on the endpoint devices.
1: https://github.com/lf-edge/glossary/blob/master/edge-glossary.md#edge-computing “Open Glossary of Edge Computing – Edge Computing”