How the Internet Actually Works

Not a Cloud: The Physical Reality

The internet is routinely described as "the cloud" — a metaphor that implies something ethereal and weightless. The reality is the opposite. The internet is a physical infrastructure of extraordinary scale: approximately 1.3 million kilometres of undersea fibre optic cables running across the ocean floor, thousands of data centres consuming gigawatts of electricity, and the routing hardware of hundreds of thousands of interconnected networks.

When you send an email from Frankfurt to someone in Tokyo, the data almost certainly travels through one or more of the 400+ undersea cable systems that form the backbone of international internet traffic. Most cross-Atlantic data travels through cables landing in the Netherlands or France. Most Trans-Pacific data travels through cables landing in California or Oregon. The cables are typically no wider than a garden hose and are occasionally severed by fishing trawlers, ship anchors, and earthquakes.

Packets, Not Circuits

The fundamental design insight of the internet is packet switching. Before the internet, telephone networks used circuit switching: a dedicated physical connection was established between caller and receiver for the duration of the call. This was reliable but inefficient — the circuit was occupied even during silences.

Packet switching works differently. Data is broken into small chunks called packets. Each packet is labelled with its origin, destination, and its position in the sequence. Packets travel independently across the network, potentially via completely different routes, and are reassembled at the destination. A single video call might have its packets routed through Chicago, Frankfurt, and Singapore before arriving in the right order.

This approach was intentionally resilient. The original ARPANET, from which the internet evolved, was partly motivated by the desire to build a communication network that could survive a nuclear strike on any one node by dynamically rerouting around damage. That design philosophy — no central point of control, decentralised routing — remains embedded in the internet's architecture today.

IP and TCP: The Core Protocols

Two protocols define how data moves across the internet. The Internet Protocol (IP) defines addressing: every device connected to the internet has an IP address (like 192.168.1.1), and every packet carries the IP address of its destination. IP handles routing — getting packets from origin to destination.

The Transmission Control Protocol (TCP) handles reliability. It ensures that packets arrive in the correct order, requests retransmission of lost packets, and confirms successful delivery. TCP adds overhead — more back-and-forth between sender and receiver — but guarantees that data arrives complete and correct. For applications where some data loss is acceptable (live video, voice calls), the lighter UDP protocol is often used instead.

DNS: The Internet's Phonebook

You navigate to websites using domain names like briefsy.net, not IP addresses. But the internet's routing system works on IP addresses, not names. The Domain Name System (DNS) bridges the gap — it is a distributed database that translates human-readable names into machine-readable IP addresses.

HTTP and HTTPS: The Web Layer

The internet and the web are often used interchangeably but are distinct. The internet is the underlying network infrastructure. The World Wide Web is one application built on top of it — a system of interlinked documents accessible via browsers.

The web runs on HTTP (Hypertext Transfer Protocol), a set of rules for requesting and transmitting web pages. When your browser connects to briefsy.net, it sends an HTTP GET request to the web server. The server responds with the HTML, CSS, and JavaScript files that make up the page. Your browser assembles these into what you see.

HTTPS — the padlock in your browser — adds encryption via TLS (Transport Layer Security). Without it, every network node your data passes through could read it in plain text. With HTTPS, the data is encrypted end-to-end between your browser and the web server, making interception by intermediaries effectively impossible in practice (though not in theory).

Why It's Fragile and Robust Simultaneously

The internet's decentralised architecture makes it remarkably robust against the failure of any single component. If a major data centre in Frankfurt goes offline, traffic reroutes around it. If a submarine cable is severed, backup routes activate automatically. No single point of failure can take down the global internet.

Yet the internet is also surprisingly fragile in practice. A significant fraction of web traffic passes through a small number of content delivery networks — Cloudflare, AWS, Akamai, Fastly. When Cloudflare experienced a major outage in 2022, thousands of major websites became inaccessible simultaneously, despite the fact that the underlying internet infrastructure was functioning perfectly. The logical centralisation of the web on a small number of infrastructure providers recreates single points of failure at the application layer even as the network layer remains distributed.