Longtime readers will have noticed my meaningful use of the word edge. I often use it in a way that implies evolution on top of what we all consider to be IoT devices. I believe the development of IoT and the development of Edge to be parallel movements - both continue forward to this day, and both are as relevant as ever. Edge isn't replacing IoT - it's augmenting it.
Rather than comparing and contrasting Edge and IoT (I do love a well-executed Venn diagram), let's briefly talk about the history of the Internet of Things.
The birth of the internet of things came by way of an (apparent) observation of computer technology: computers are reliable. Computers don't need coffee breaks, and they don't tire after a long day. However, machines are very predictable - they do precisely the tasks asked of them: no more / no less. For mundane tasks, this is perfect! Write a computer program to alert you when CNN adds a new story - Easy, the computer will happily sit for hours/days/weeks watching CNN for a news story, and the moment one is posted: it will alert you.
Consider a basic IoT example: a grocery store with many refrigerators. As it stands now, the owner must send employees at regular intervals to take temperature readings. A non trivial amount of trust (and reliability) comes into play here as the owner relies on the employee. When the employee fails, the owner risks a citation and spoiled food.
What should our savvy owner do?
Our savvy owner decides to convert her store into an IoT smart store (yes, smart is marketing lingo...)! During the store's off-hours, our owner places temperature sensors into each refrigerator. These sensors are incredibly bare: they consist of a sensor, Wi-Fi chip, and small microcontroller (similar to a CPU). The sensor devices are simple: take a temperature reading every second and send the data to a control computer in the back of the store (or on the cloud, but let's keep this story simple).
Initially, the owner may have her employee's work in shifts watching a computer screen as temperature readings stream by monitoring the situation. The new IoT installation is already a step better than the initial strategy: no accidental walking past a refrigerator. Now the owner goes a step further: a program monitors the temperature readings and sounds an alert over the store's intercom when a refrigerator requires attention. She's in an improved place: no missed refrigerators and no missed alerts!
Act I. Scene.
The store's refrigerator monitors are a classic IoT setup. Our owner is doing no in-store processing of the temperature data — no attempt is being made to learn trends. As a result, a straightforward (manual) task is now an automated task - one could even say that her refrigerators are smart as they will alert her when they need attention.
The local power company visited our enterprising owner with a request: lower your power consumption by 10% during peak hours, and they will grant her credit on her power account. Our owner has all of the data that she needs to know if she should agree to this: temperature readings and power readings from her meter. Now she wants trends/analytics, and perhaps a machine learning model.
Will her classic IoT setup provide the answers that she seeks?
Why do I like refrigerator examples?
Part 2 has (at least) one of the answers.