ISO/OSI layer

When two computers exchange data, a great many things have to work at the same time. A cable or radio signal must be present. The data must find the right recipient. And in the end, a program must also understand the received characters. So that a single, gigantic set of rules doesn’t have to describe all of this at once, the standards organization ISO broke these tasks down into seven stages. Each of these stages is called a layer, and the overall model is called the ISO/OSI model.

Why seven stages instead of one overall solution

The benefit lies in the division of labor. Each layer is allowed to assume that the layer below it does its job. The level responsible for routing in the network therefore doesn’t need to know whether the data travels over copper cable, fiber optics, or WLAN. Conversely, the cable doesn’t care whether a video or an email is currently being transmitted.

This means individual parts can be swapped out without touching the rest. Switching from WLAN to mobile data only changes the lower layers. The browser on top notices nothing of this and keeps loading the page. This separation is precisely the reason the internet has grown over decades without having to be reinvented.

A second advantage is the shared language. When a technician says the problem lies on layer 3, every expert immediately knows that it concerns addressing and routing, not a broken cable. So the model is also an organizing framework for troubleshooting.

What happens at the individual levels

At the very bottom is layer 1, the physical layer. It describes connectors, voltages, and radio frequencies — in other words, the pure transmission of zeros and ones. Layer 2 ensures that devices on the same local network can reach each other directly and detects transmission errors. Layer 3 handles addresses and routes across many networks; this is where IP addresses and routers operate. Layer 4 ensures that all parts of a transmission arrive completely and in the correct order.

The upper three levels are closer to the application. Layer 5 manages sessions, i.e., the temporal framework of a conversation between two programs. Layer 6 handles presentation, such as character sets or encryption. Layer 7 is the application layer, with the rules by which browsers, mail programs, or apps communicate.

A helpful analogy is a letter within a company. The text is the application layer. The envelope with the address corresponds to routing. The mail carrier and his bicycle are the physical level. Each station adds something around the message when sending, and the recipient unpacks it again in reverse order.

From textbook to everyday networking practice

In reality, the internet doesn’t work exactly according to these seven layers. It uses the older and leaner TCP/IP model with four levels. Nevertheless, the ISO/OSI model has persisted because it subdivides things more finely and is well suited for explanation. Anyone training in IT learns it in almost every curriculum.

In everyday professional life, one constantly encounters the layer numbers. A switch is called a layer 2 device, a router operates on layer 3. Firewalls are distinguished by whether they only check addresses or look all the way into the application layer. Attacks, too, are categorized this way, such as when people speak of a denial-of-service attack on layer 7.

A common misconception is the assumption that every network device belongs to exactly one layer. Modern devices often cover several levels. The model describes tasks, not components. Read this way, despite its age, it remains a very useful tool for understanding networks.