The one-sentence version
The internet is billions of computers connected by cables and radio waves, all agreeing to speak the same language so any one of them can send data to any other.
No company owns it. There's no central computer running it. It's closer to the global postal system than to a single machine: lots of independent carriers, one shared addressing scheme, agreed-upon rules for handing parcels along.
The postal system analogy
Keep this analogy — it maps almost perfectly:
| Postal system | Internet | Technical name |
|---|---|---|
| House address | A computer's numeric address | IP address (e.g. 142.250.77.46) |
| Phone book: name → address | Domain name → IP address | DNS |
| A parcel | A small chunk of data | Packet |
| Sorting offices that route parcels | Machines that forward packets hop by hop | Routers |
| Rules for packaging & delivery confirmation | Rules for splitting, sending, confirming data | Protocols (TCP/IP) |
IP addresses — every machine needs an address
An IP address (Internet Protocol address) is a number that identifies a
machine on a network, like 142.250.77.46. If a computer wants to receive
data, it needs one — exactly like a house needs an address to receive mail.
Your laptop gets one from your Wi-Fi router; your router gets a public one from your ISP (Internet Service Provider — the company you pay for internet, like Jio or Comcast). Servers that host websites have stable, well-known addresses.
DNS — the internet's phone book
Humans are bad at remembering 142.250.77.46, so we use names:
google.com. DNS (Domain Name System) is the worldwide directory that
translates names into IP addresses.
When you type google.com, your computer first asks a DNS server "what's the
IP for google.com?", gets back the number, and then makes the real
connection — the same way you look up a friend's address before mailing a
letter.
DNS is one of the most common real-world failure points — "is it DNS?" is a running joke among engineers. It's also an interview favorite: "What happens when you type google.com and press Enter?" is asked at every level from intern to senior. This page is that answer, in beginner form.
Packets — data travels in small pieces
The internet doesn't send your photo as one big blob. It chops everything into packets — chunks of roughly a thousand bytes, each labeled with the sender's address, the destination address and a sequence number, like numbered parcels.
Why bother?
- Sharing — small packets from millions of users interleave on the same cables, so nobody hogs the line.
- Resilience — if one packet is lost or one route dies, just that packet is re-sent, possibly along a different path.
- No single path needed — each packet can take whatever route is open, and the receiver reassembles them in order.
Routers are the sorting offices: each one reads a packet's destination and forwards it one hop closer. A packet from Mumbai to a server in Virginia might pass through 15–20 routers and an undersea fiber-optic cable, in about 200 milliseconds.
Protocols — the agreed-upon rules
A protocol is just a rulebook both sides agree to follow. The internet is a stack of them, each solving one problem:
- IP (Internet Protocol) — addressing and routing: where does this packet go?
- TCP (Transmission Control Protocol) — reliability: numbers the packets, confirms receipt, re-sends lost ones, reassembles in order. Like a courier with delivery confirmation.
- HTTP (HyperText Transfer Protocol) — the language browsers and websites speak on top of TCP: "GET me this page", "here it is". HTTPS is the same conversation inside an encrypted envelope so nobody along the route can read it.
What happens when you type google.com and press Enter
The classic question, end to end:
- DNS lookup — the browser asks DNS for
google.com's IP address (checking its own cache first). - Connection — it opens a TCP connection to that IP and performs the HTTPS encryption handshake.
- Request — it sends an HTTP request:
GET / HTTP/1.1 Host: google.com. - Routing — the request, split into packets, hops router-to-router to a Google data center.
- Response — Google's server builds the page and streams it back as packets, which your machine reassembles.
- Rendering — the browser turns the response (HTML, CSS, JavaScript — next page) into pixels on your screen.
Total time: typically 100–500 ms. Each of these six steps becomes a whole field of engineering later in the roadmap.
Industry perspective
- Google/Netflix/Meta place servers physically close to users (CDNs — Content Delivery Networks, covered in Level 6) because even at the speed of light, distance costs milliseconds, and milliseconds cost users.
- WhatsApp keeps a single long-lived connection open per phone instead of reconnecting per message — that's why messages feel instant.
- Every outage headline you read ("X is down") is one of the layers above failing: DNS misconfigured, routes withdrawn, servers overloaded.
Common beginner mistakes
- "The internet and the web are the same thing." The internet is the network (roads); the World Wide Web is one service running on it (websites, via HTTP). Email, video calls and online games use the internet but aren't the web.
- "Wi-Fi is the internet." Wi-Fi is just the wireless link from your device to your router — the last 10 meters of a journey that's thousands of kilometers of fiber.
- Thinking data goes directly from A to B. It hops through many routers, ISPs and cables owned by different organizations, none of which see the whole picture.
Interview perspective
Check yourself
- Your friend's website is unreachable, but you can still reach google.com. Using the postal analogy, list three different places the failure could be.
- What's the difference between your Wi-Fi, your ISP and the internet?
- Why does a video call to a nearby city feel snappier than one to another continent, even on fast connections?
Next: Websites & Browsers — what's actually inside the pages that travel over these wires.