The OSI model splits the job of sending data between two computers into seven layers, each with one responsibility. Engineers and cloud docs use its layer numbers to say where something happens, so knowing them explains a lot of networking vocabulary.
Why networking is split into layers
Sending 'Bring tuna' from your laptop to a friend's phone involves a lot of separate problems. The app has to produce the message. Something has to agree on a format both ends understand. Something has to make sure nothing gets lost, something has to find a route across the internet, and something has to move electrical signals or radio waves from one device to the next.
Putting all of that in one piece of software would make every part hard to change. Layering solves this by giving each problem to one layer, which only talks to the layer directly above and below it. Each layer relies on the one below to do its job and doesn't care how it does it.
That is why your browser works the same over Wi-Fi, Ethernet or 5G. The browser lives at the top. The way bits physically move lives at the bottom, and it can be swapped without the browser noticing.
The OSI (Open Systems Interconnection) model is a reference model, published as a standard by ISO and the ITU. It describes what each layer is responsible for. It doesn't say which protocols you must use.
The seven layers
The layers are numbered from the bottom, but people usually walk through them from the top, because that is where your message starts.
| Layer | What it does | Examples |
|---|---|---|
| 7 Application | What the user's software speaks | HTTP, DNS, SMTP |
| 6 Presentation | Formats, encrypts, compresses | Character encodings, TLS |
Wondering how this was made? How we make these