What Happens When You Type a URL and Press Enter?

Here is the full journey, from pressing Enter to seeing the page:

Let's go through each step.

Step 1: The browser reads the URL

A URL (Uniform Resource Locator) is the full address of a page or file on the web. A URL like https://www.amazon.com/deals?page=2 has parts:

• https is the protocol: the set of rules for how to talk (secure HTTP).
• www.amazon.com is the domain name: the human-friendly name of who to talk to.
• /deals is the path: what you want.
• ?page=2 is the query string: extra details.

If you typed something that is not a URL, like "best laptops", the browser sends it to a search engine instead.

Step 2: DNS, the internet's phone book

Your computer cannot connect to "www.amazon.com". It needs a number, an IP address. An IP address is the number that identifies one machine on a network. DNS (Domain Name System) is a worldwide system of servers that turns a name into that number. A cache is a saved copy of an answer, kept so that the next request is faster.

The lookup tries the fastest options first:

1. Browser cache. Did I look this up in the last few minutes?
2. Operating system cache. Did any app on this computer look it up?
3. Recursive resolver. A DNS server that looks up the answer for you. It is usually your internet provider's DNS server, or a public one like Google's 8.8.8.8 or Cloudflare's 1.1.1.1. It does the hard work for you.

If the resolver doesn't know the answer, it asks a chain of servers. The root serversknow who runs each top-level domain, such as .com. The TLD servers know who runs each name under that top-level domain, such as amazon.com. The authoritative server (Amazon's own DNS here) holds the final answer.

The resolver saves (caches) the answer for a while and gives it to your computer.

TTL (Time To Live). Every DNS answer comes with a TTL. It is a number of seconds that says how long the answer may be cached. This creates a trade-off:

• Long TTL (for example, a day): fewer lookups and faster browsing. But if the company changes its server IP, some users keep going to the old address for up to a day.
• Short TTL (for example, 60 seconds): changes spread quickly, but the DNS servers have to answer more queries.

That is why engineers lower the TTL a few days before a server migration.

Common DNS record types. A DNS record is one saved fact about a domain. Each record has a type:

When you connect your own domain to a hosting platform like Vercel, you usually add an A or CNAME record. This means you are editing DNS.

Step 3: IP addresses, packets and routers

Now the browser knows the IP address, so it can start talking.

IP addresses come in two versions:

• IPv4 looks like 192.168.1.10. There are only about 4.3 billion of these, and the free ones have run out.
• IPv6 looks like 2001:db8::1. There are so many that we will not run out.

Public vs private IPs. A public IP can be reached from the whole internet. A private IP works only inside one local network. Your laptop at home probably has a private address like 192.168.1.5. Your Wi-Fi router has one public address that every device in the house shares. The router uses NAT (Network Address Translation). NAT is a trick where the router rewrites outgoing packets so they carry the public address, and remembers which reply belongs to which device. This is why your laptop can visit websites, but a stranger on the internet cannot connect directly to your laptop.

Packets. Your request is not sent as one big piece. It is cut into packets, usually up to about 1,500 bytes each. Each packet carries:

• a source IP (you),
• a destination IP (Amazon),
• and a piece of the data.

Routers are the sorting centres. A router is a device that passes packets from one network to the next. Each router looks at the destination IP and sends the packet one "hop" (one step) closer. A router does not read your data. It only looks at the address on the envelope. A packet from India to a US server might pass through 10 to 20 routers.

You can see this yourself with a command-line tool that lists every router on the path:

traceroute www.google.com     # macOS / Linuxtracert www.google.com        # Windows

Each line is one router on the path, with the time it took to reach it.

How do routers know the way? Big networks (internet providers, cloud companies, Google, Facebook) use BGP (Border Gateway Protocol). BGP is the protocol that networks use to announce "I can reach these IP addresses". Other networks learn routes from these announcements. BGP is the internet's map, built by many companies telling each other which roads they own.

Ports. An IP address finds the machine. A port is a number (0 to 65535) that finds one program on that machine, like an apartment number inside a building. Some well-known ports:

• 80 for HTTP (the web protocol, without encryption),
• 443 for HTTPS,
• 22 for SSH (a protocol for logging in to a remote machine safely),
• 5432 for PostgreSQL (a database),
• 53 for DNS.

Step 4: The network layers

Networking is organised in layers. A layer is one level that does one job and trusts the layer below it. The simple TCP/IP model has four:

You will also hear about the OSI model, which has 7 layers. In system design, people often say "Layer 4" and "Layer 7". These numbers come from the OSI model:

• Layer 4 means the transport layer (TCP and UDP, the two ways to send data between programs).
• Layer 7 means the application layer (HTTP).

This matters later. A load balancer is a server that shares incoming requests between many identical servers. An L4 load balancer only sees IPs and ports. An L7 load balancer can read the URL and headers.

Step 5: Connect, secure, request

Next, the browser:

1. opens a TCP connection (a quick three-message "hello" called a handshake, covered in the next lesson),
2. sets up TLS encryption for HTTPS (lesson 3). TLS (Transport Layer Security) is the protocol that scrambles data so that nobody on the path can read it,
3. sends an HTTP request (HTTP is the protocol for asking a server for a page), such as GET /deals?page=2.

Step 6: The server does the work

The request usually does not go to a single computer. On a big site it passes through several tiers. A CDN (Content Delivery Network) is a group of servers around the world that keep copies of files close to users. A cache such as Redis is a very fast store that keeps recent answers in memory. The database stores the real data for the long term.

Each of these gets its own lesson later in the course.

Step 7 and 8: Response and rendering

The server sends back an HTTP response, like 200 OK with HTML. HTML is the text that describes the page. The browser:

1. reads the HTML,
2. finds links to CSS, JavaScript and images,
3. requests all of those too (often dozens of extra requests),
4. builds the page on your screen.

Open your browser's DevTools, go to the Network tab and reload any site. You will see every request, and how much time each one spent on DNS, connecting, TLS, waiting and downloading.

Why distance matters: latency

Latency is the time it takes for data to get from one place to another. A round trip is the trip there and back. Signals in fibre-optic cable travel at about 200,000 km per second, roughly two-thirds of the speed of light. That sounds instant, but:

• London → New York is about 5,500 km. One way takes at least about 28 ms, so a round trip takes about 56 ms.
• London → Sydney is about 17,000 km. The round trip is at least about 170 ms, and often 250 ms or more in practice.

A page that needs 4 or 5 round trips before showing anything will feel slow to someone far away, no matter how fast the server is. This is why companies put servers and CDNs close to users. You cannot make light faster, but you can make the distance shorter.