How networks work

Why networks exist

Devices used to be islands: you carried a file on a floppy disk and a printer served just one computer. A network is two or more devices that can exchange data. It solves sharing – of files, a printer, one internet connection, and messages. Today almost everything is networked: your phone, laptop, TV, game console, smart speaker. The principle is always the same – connect devices so they can talk and share.

How a message is split into packets

How does a file actually travel across the network? It is not sent as one big blob. It is split into many small chunks – packets. Each packet is like a letter: the content (your data) inside an envelope marked to and from. The network does not care what is inside – it reads the address on the envelope and delivers accordingly. At the destination the packets are put back together in the right order. Step through it (‹ / ›).

The router and the internet

That box from your provider at home is the router. It does several things, but its main job is one: decide what stays home (your LAN) and what goes out to the internet (the WAN). And how does a packet reach a remote server? One router hands it to the next, and that one to the next – like changing from train to train. The internet is exactly that: a network of networks, routers worldwide linked together, owned as a whole by no one. Step through the packet’s journey.

Bits and signals

Whether data travels by cable or through the air, at the very lowest level it is always just ones and zeros – bits. But how does a bit physically travel from one device to another? It depends on the medium – what it travels on. Copper carries a bit as voltage, fiber as light, and Wi-Fi as a radio wave. How many bits a medium carries per second is its bandwidth – measured in Mbps or Gbps.

Wired media: copper and fiber

There are three main types for a wired connection. Twisted pair (the RJ-45 plug) is the most common cable in home and company networks – usually the unshielded UTP variant, with shielded STP/FTP for noisy environments (near power cables, motors or machines that induce interference into the wire). Optical fiber guides light and offers higher speeds and longer reach – it forms the backbone of the internet. Coax is older, today mostly cable TV and cable-company internet. Practical tip: where you can, use a cable – it is steadier and faster than Wi-Fi.

Wireless media

There are mainly two ways to connect wirelessly. Wi-Fi (the 802.11 standard) is a radio link to an access point (at home it is part of the router). Mobile networks (4G/5G) are a completely different infrastructure – the operator’s masts cover whole regions; 5G adds higher speed and lower latency. And there are short-range wireless networks too, such as Bluetooth (headphones, mouse, smartwatch). A comparison of the main ones:

Why we need addresses

A network can easily have thousands of devices. How does each know who to send data to? Just like the postal system: every device needs an address. But in networking one is not enough – there are two, each for a different job. One says who you are (your specific device), the other where you are (which network). In the next modules we look at both – first "who you are" (MAC), then "where you are" (IP).

MAC address: the factory label

The first address says who you are. It is the MAC address – like a network card’s factory label. The maker "burns" it in at manufacture and it normally never changes. It looks like six pairs in hexadecimal, e.g. 00:1A:2B:3C:4D:5E (each pair = 8 bits, 48 in total). It is unique worldwide (no two cards share one), but it is used only within one local network – switches (the boxes that carry data around inside a network) deliver by it only inside your network, and it is not carried across a router. The moment data leaves your network, the second address, IP, takes over (in the next module).

IP address: the postal address

An IP address solves what MAC cannot: it is hierarchical – it has a network part (picture it as a city) and a host part (like a house number in that city). That lets a router tell which direction an address lies. IPv4 is written as 4 numbers 0–255 (e.g. 192.168.1.10); under the hood it is 32 bits (a bit = the smallest piece of data, a 1 or 0 – you met bits in the media lesson). Not every address goes out to the internet: some ranges are private – valid only inside your home or company network and never routed on the internet: 10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16. Public addresses are globally unique and assigned by your provider (ISP); you need one for the internet to reach you. The /24 notation says how many leading bits belong to the network – in 192.168.1.10/24 that is the first three numbers (192.168.1 = network, .10 = host); we cover subnetting in the next course.

Your router has two addresses

At home every device has a private IP (10.x.x.x, 192.168.x.x, 172.16–31.x.x) – these work only inside your network and the internet could never find them. But the router also has one public IP from your provider – that one the whole internet sees. When you send data out, the router swaps your private address for its public one (and back on the reply). That translation is called NAT. Thanks to it the whole household hides behind a single public address. Step through the swap. How NAT works in detail comes in the next course.

ARP: finding the MAC for an IP

You want to send data to someone on your local network and you know their IP address. But inside the network a frame is physically delivered by MAC address (the network card’s factory label), not by IP – without the target’s MAC you have no one to physically hand the data to. How do you find that MAC for a given IP? That is exactly what ARP solves – step through it.

Default gateway: the way out

What if the target is not on your network – say a website somewhere on the internet? You cannot reach it directly. So you send the data to the default gateway: that is your router, the "way out" of the network. In practice the gateway is usually the first address in the network, e.g. 192.168.1.1. Your device doesn’t have to guess it – it gets it automatically together with its IP address (from DHCP, in the next module). The rule is simple: target on my network → deliver directly; target outside → send to the gateway and it forwards it on. (The device tells whether a target is "on my network" from the network part of the IP address – if it matches its own, the target is local.)

DHCP: your automatic address

Where does the IP address actually come from? Usually you do not set it by hand – a DHCP server assigns it automatically (at home it is in the router). The assignment is a short conversation of four messages – step through it. That is why joining the Wi-Fi just works. If DHCP fails, the device gives itself a fallback address 169.254.x.x (called link-local) – but that cannot reach out, so there is no internet.

When it does not work: first steps

When the internet is down, a few first steps usually help – based on the symptom. 1) No internet but Wi-Fi is up: you probably have no valid IP address (maybe DHCP did not give one) – try disconnecting and reconnecting. 2) One site fails, others work: the issue is usually DNS (the name-resolution service – covered in the next lesson) or a stale DNS cache – the site itself could also be down. 3) Nothing works at all: try a ping to the default gateway (e.g. 192.168.1.1). A ping is a simple "are you there?" test – in the command line (the system’s text window where you type commands instead of clicking) you type ping 192.168.1.1 and wait for a reply. If the gateway does not answer, the problem is on your local network (cable, network gear, Wi-Fi), not out on the internet.

DNS: the internet's phone book

In your browser you type names like example.com – easy for you to remember. But computers and networks only use numbers between themselves, IP addresses. Something has to translate that name into a number – and that is exactly what DNS does: the internet’s phone book. Step through what happens when you open a page. If DNS fails, the page will not load even though its server is up – the browser simply cannot find the way to it.

TCP and UDP: reliable vs. fast

You already have the address (IP) and the path (via the router) – data reaches the right device. But many apps run on it at once: a browser, email, a game. How does the data get to the right app (the one waiting for it), and reliably at that? There are two approaches. TCP is like registered mail: the two sides first agree they are talking, then confirm the transfer to each other – anything lost is resent. So it is reliable (web, files, email) but a bit slower. UDP is like a postcard: it just sends and does not wait for confirmation. It is fast but with no guarantee – ideal for live calls, video and games, where resending something that should have arrived long ago is pointless. We’ll show exactly how that "agreement" works in the next course.

Ports: the apartment number

One computer runs many apps at once – browser, music, email – all sharing one IP address. How does data know which app it belongs to? By the port. If the IP address is the building, the port is the apartment number. Port numbers range from 0 to 65535, and in a conversation each side has one: the server listens on a pre-agreed number so the client can find it. Common services have standardized numbers (the well-known ports, 0–1023) – 80 HTTP, 443 HTTPS, 53 DNS, 22 SSH – but a server can listen on any other port too (say 4070 or 8080) as long as the client knows it. The client in turn takes a random high port (an ephemeral one = temporary; the system assigns it just for this one connection). A connection is then pinned down by four things (plus whether it’s TCP or UDP): your IP + your port and the server’s IP + the server’s port. That is why replies never get mixed up, even with many connections open at once.

HTTP: how the browser talks to the server

Once you have the server’s address and a connection, how does the web actually "talk"? In the HTTP language. It works as request → response: the browser (client) sends a request – a method (e.g. GET = give me, POST = here is data), the page address and headers. The server returns a response: a status code (200 OK, 404 Not Found, 500 server error) and a body – e.g. the HTML page. It is all plain text, so anyone on the path could read it. That is why HTTPS exists (in the next module), encrypting the content.

HTTPS and encryption: why the padlock

What you send over the network does not go straight to the target – it passes through plenty of machines that are not yours (the café Wi-Fi, your provider, the routers along the way). HTTP is the language of the web: you send a request, the server returns a response (the page). But it sends everything as plain text – whatever you send (passwords, card numbers) can be read by anyone on the way. HTTPS = HTTP + encryption. It encrypts the data so that only the two ends can unlock it – you and the server. To anyone on the path it is an unreadable jumble, and the padlock shows in your browser. Note: HTTPS hides the content, but not which site you connect to (the server name is visible).

What HTTPS does not hide

HTTPS encrypts the content – passwords, card numbers, message text. That is great, but it is worth knowing what it does not hide. Think of a parcel: the contents are sealed, but the address on the outside is visible. An observer on the path (say your provider) can see which domain you connect to (the server name, SNI), when and for how long and how much data flows – just not what exactly you exchange. Example: it is visible that you connected to banka.cz, but not what you do there. That is why name resolution is also being encrypted – DNS over HTTPS (DoH); more on that elsewhere.

From click to loaded page

Now you have all the pieces. Let us walk through what happens when you type an address into the browser and press Enter – no new terms, just putting it together into one journey.

Most of the steps happen in under a second, a hundred times a day.

What now?

That is it! You now know how data travels from your computer over Wi-Fi or a cable, through the router, DNS, TCP and HTTPS to the server and back – and the page loads in a fraction of a second. But it does not end there: What if someone on that path were eavesdropping? And how far does an attacker who gets into the network reach? All those devices, protocols and "floors" have a system and names – ISO/OSI and TCP/IP. That is the next course: layer by layer, device by device, until you can name every piece of the journey and understand where a network can be secured.