Edge Computing
February 26, 2024
Overview
Edge Computing can be seen as a natural and ongoing expansion of the location of processing from centralized data centers... to you. The purpose of this expansion is to decrease the latency of requests between your devices and these locations of processing. It also increases utilization of dispersed processing capabilities.
Table of Contents
What is Considered Edge Computing
Edge Computing is more than just setting up a data center near end-users. For example, if someone had a local brick-and-mortar business, and they decided to distribute their website from a data center that is located in the same city as their customers, this is not technically Edge Computing. Edge Computing requires some level of redistribution or outsourcing of processing from one or more centralized locations. The classical examples of Edge Computing are Edge Data Centers.
Edge Data Centers are similar to Content Delivery Networks (CDNs), but they provide compute (processing) capabilities as well. CDNs are locations that cache data closer to the user.
On The Fence 'of Edge Computing'
As mentioned in the Overview, Edge Computing is an ongoing expansion of the location of processing from centralized data centers. This ongoing expansion is not just more Edge Data Center locations, it is the opinion of Software.Land that also IoT devices will one day be considered Edge Computing.
The definition of IoT is open to interpretation. However, the key distinction to Edge Computing is the redistribution or outsourcing of processing out of one or more centralized locations. If the IoT device is engaged in local processing of data, then it can be considered part of the Edge Computing ecosystem. It is important to note that this is not a traditionally accepted point of view.
Content Delivery Networks (CDNs) are not considered Edge Computing in the absolute strictest definition (many do consider them to be as a matter of tradition), however, they are considered to be part of the Edge Computing ecosystem. They don't engage in any significant compute, but they do shorten the distance from user devices to data. One could also make the argument that there are still servers inside CDNs that must engage in some compute and processing in order to fetch that cached data. We can furthermore extend this Edge Computing ecosystem definition to include IoT devices that strictly cache data. Although, this is not widely accepted.
Alternate Definitions
Edge Computing is entirely customer-centric. It's strictly about shortening the distance from user devices to their compute location (and data location, if we're considering the Edge Computing ecosystem). If that distance can be shortened to zero (all computing takes place in the user's device), all the better. Static websites, like Software.Land's blog, (see Static vs Dynamic Website) are not considered to be valid examples of the Edge Computing ecosystem, but it's the opinion of Software.Land that they should be. The website's data is cached locally, resulting in what can be considered an edge caching location directly in the device. Standard (dynamic) websites also engage in some caching, allowing another extended definition of the Edge Computing ecosystem. However, neither of these extended definitions are widely accepted.
The Future
As mobile network bandwidth increases, as latencies decrease, and as compute power and storage become more dense over time, the boundary of Edge Computing pushes forward towards you. The possibilities of what can be accomplished nearer to you in smaller form factors increases. This increase is a never-ending upward trajectory. This frontier will continue to move outward from centralized data centers to you, helping to reduce tediousness and increase productivity.
For a more traditional definition of Edge Computing, see Wikipedia's article.