- Stage
- Now · 2026 to 2027
- Area of work
- Reinnder Cloud & Software
- Used in
- Factories · Shops and buildings
Figure · Illustration
From the device to the cloud
-
A sensor, camera or machine makes the data and can make the very quickest decisions itself.
-
A small computer in the same room or building combines many devices and decides within milliseconds.
-
A server at a telephone exchange or cell tower serves a whole neighbourhood, a little further away.
-
A large data centre, possibly in another country, stores history, trains models and coordinates many sites.
What the two words mean#
The US standards body NIST defines cloud computing as a model for on-demand network access to a shared pool of configurable computing resources, such as servers, storage and applications, that can be provisioned and released quickly with little effort from the provider. In plain terms, you rent computing from someone who runs very large buildings of machines, and you pay for what you use.
Edge computing moves some of that work to the other end: onto the device itself, onto a small computer beside it, or into a nearby site such as a factory room, a cell tower or a shop. The word “edge” means the edge of the network, where the data is created. NIST has also described a middle layer, fog computing, which spreads applications, management and analysis across the network between the devices and the cloud.
Why distance matters: a worked example#
Signals in an optical fibre travel at roughly two-thirds of the speed of light in a vacuum, which is about 200 kilometres every millisecond. A data centre 1,000 kilometres away therefore adds at least 10 milliseconds to every round trip, from the glass alone, before any switch, queue or computation is counted. No software can remove that delay; only moving the work closer can.
For a web page, ten milliseconds is invisible. For a robot arm that must stop when a hand enters its path, or a machine that must reject a faulty part on a moving belt, the delay and its unevenness can matter. Distance also costs data: a camera can produce far more video than is worth sending, so a nearby computer can look at it and send on only a short report.
How the two are combined in practice#
A common pattern is to react at the edge and learn in the cloud. A local computer makes the quick decisions and keeps working when the link is down; the cloud collects summaries from many sites, trains better models, stores history and sends updates back out. The device does what cannot wait, and the cloud does what needs scale.
The price of the edge is that you now look after many small computers in many places instead of a few big ones. Updates, security patches, spare parts and monitoring all multiply. That is why the cloud stays the default, and the edge is chosen for a reason: speed, privacy, cost of data, or independence from the network.
Where it is used#
- Factories Checking parts on a moving line and stopping a machine within moments, with summaries sent on for planning.
- Shops and buildings Counting stock or people and controlling heating on site, even when the internet connection drops.
- Vehicles and robots Deciding what to do next on board, and uploading what they learned later, when a good link is available.
- Websites and apps Serving pages and files from servers near the user, while the main systems stay in a data centre.
When should the work stay in the cloud?#
Keep work in the cloud when it needs a lot of computing for a short time, when it combines data from many places, or when nobody is waiting for the answer. Training a model, running a monthly report, storing years of history and serving a website to people around the world all suit the cloud, because they gain from size and from being managed in one place.
The cloud is also the right place to start. Moving work to the edge adds machines to maintain, so it is wise to prove the idea in the cloud first and move only the parts that have a reason to be close.
Edge compared with cloud, and with doing it all on the device#
The cloud offers almost unlimited, rentable computing and is easy to manage from one place, but it is a long way off and depends on the link. The edge is near and keeps working when the link fails, but each site has limited computing power and has to be looked after.
Doing everything on the device is the extreme case of the edge: fastest and most private, but limited by the device’s size, battery and cost. Most designs choose a middle path, with the device, a nearby computer and the cloud each doing the jobs they suit.
| Compared on | On the device | Nearby edge | Cloud |
|---|---|---|---|
| Distance from the data | None | Metres to a few kilometres | Often hundreds of kilometres or more |
| Reaction time | Fastest | Fast and steady | Slower, and varies with the network |
| Works when the link fails | Yes | Yes, for what it holds | No, unless something local covers for it |
| Computing power | Smallest | Moderate | Very large, rented as needed |
| Looking after it | Many separate devices | Many separate sites | One place, run by the provider |
What goes wrong at the edge?#
The common problems are not about speed. A small computer in a cupboard has no one to restart it, an update that fails on one site in fifty leaves the fleet in two versions, and a device in a public place can be stolen or tampered with. Power cuts, heat and dust are ordinary events, not rare ones.
Teams that succeed plan for these from the start: updates that can be rolled back, a way to see whether every site is healthy, encryption of stored data, and a clear rule for what the site does when it cannot reach the cloud at all.
Key terms#
- Latency
- The delay between asking for something and getting the answer, usually counted in milliseconds.
- Round trip
- The journey of a message to a computer and of the reply back again; distance sets a minimum time for it.
- Edge node
- A computer placed close to where data is created, such as on a machine, in a building or at a cell tower.
- Fog computing
- A layer of computing between devices and the cloud that spreads applications and analysis across the network.
Common questions#
Will edge computing replace the cloud?
- No. They do different jobs and usually work together. The edge reacts locally; the cloud stores, trains and coordinates across many sites.
Is edge computing more secure than the cloud?
- Not automatically. Keeping data on site can reduce what travels and who sees it, but many small devices are harder to patch and to guard than a few well-run data centres. Security depends on how each is managed.
Does edge computing need 5G?
- No. It can run over Wi-Fi, cable, cellular or no network at all for a while. Fast mobile networks can make a nearby site easier to reach, but they are not required.
Sources and further reading#
Independent pages we checked while writing this guide. They are not Reinnder products, and Reinnder is not affiliated with them.
Reinnder’s angle
How Reinnder looks at edge or cloud?
Reinnder Cloud & Software is planned to design systems that use both, and to say plainly which work belongs where.
This guide explains the technology in general terms. It is not advice, and it does not describe a Reinnder product on sale.