What edge computing is
In short: Edge computing means processing data close to where it is created, such as in a store, factory, vehicle or mobile network site, instead of sending everything to a distant data centre. It reduces delay and bandwidth, and keeps working when links fail. The price is many small sites that are harder to manage and secure.
The idea in one sentence
Edge computing moves processing and storage closer to the people and devices that produce and use data. Instead of sending every camera frame or sensor reading to a central cloud, you run software nearby and send only results, summaries or exceptions.
Imagine a supermarket with fifty cameras counting queue lengths. Uploading all video would need huge bandwidth. An edge server in the store can analyse the video locally and send a few numbers each minute to head office.
Why it matters
Physics sets a limit on latency: light in optical fibre travels roughly 200 km per millisecond, and real paths add routing, queuing and processing delay. If a robot arm must react within a few milliseconds, the decision has to be made close by. Other drivers include saving bandwidth, keeping working when the WAN link fails, and laws or policies that require some data to stay in a country or on site.
Edge locations
| Location | Example hardware | Typical job |
|---|---|---|
| Device edge | Smart camera, microcontroller, vehicle computer | Filter data, run a small model, act instantly |
| On-premises edge | Small server or gateway in a shop, branch or factory | Local apps, analytics, protocol translation |
| Network or telco edge | Servers in operator sites near mobile networks | Low-latency services for many users; see MEC |
| Regional cloud | Cloud provider local zones and nearby regions | Lower-latency cloud services with familiar tools |
The term "edge" means different things to different vendors. A telecoms engineer may mean an operator site; a factory engineer may mean a box on the shop floor. Always ask "which edge?"
Trade-offs and common mistakes
- Scale of management: one data centre becomes hundreds of small sites, each needing updates, monitoring and backups. Automation is essential; see managing edge sites.
- Physical security: edge boxes sit in cupboards, poles and vehicles where anyone might reach them.
- Limited resources: less power, cooling and space than a data centre.
- Data consistency: local copies must sync with central systems after outages.
Tip: a common mistake is to move workloads to the edge without a clear driver. If latency, bandwidth, resilience or data location do not matter, central cloud is usually simpler and cheaper.
Practice
A factory needs a safety system that reacts within a few milliseconds. Should the decision run in a distant cloud region?
Give two benefits of analysing CCTV video at the edge.
What is a key operational downside of edge computing?
Key takeaways
- Edge computing processes data near where it is created.
- Main drivers: latency, bandwidth, resilience and data location rules.
- The edge can be a device, an on-site server, an operator site or a nearby cloud zone.
- Many small sites make management and physical security harder.
- Use edge only when a clear driver exists.
Original content written by Mohamed Fayas for FM TecHub. © 2026 FM TecHub. You're welcome to share a link to this page; please don't copy the text elsewhere without permission.