How the Internet Sends Data Around the World

When you send a message, open a web page or make a video call, information travels between your device and a distant computer that may be on the other side of the planet. It is tempting to imagine this information flowing along a single unbroken thread, like water through a hose. In fact the internet works in a quite different and rather surprising way. Almost everything you send is first chopped into many small pieces, and each piece makes its own journey across the network before the pieces are reassembled at the far end. Understanding this simple idea explains much of how the modern internet manages to be so fast, so flexible and so difficult to bring down.

The small pieces into which data is divided are called packets. When a file is sent, the sending computer breaks it into numerous packets, each carrying a portion of the original information. Alongside that information, every packet also carries a kind of label, sometimes called a header. The header records where the packet came from, where it is going, and which part of the whole message it represents. This last detail is important, because it allows the pieces to be put back together in the correct order even if they arrive out of sequence. A packet is therefore not just a fragment of data but a small, self-describing parcel that carries the instructions needed for its own delivery.

Every device connected to the internet is given a numerical label of its own, known as an IP address, which functions rather like a postal address. When a packet is created, the address of the destination is written into its header. The packet is then passed to specialised machines called routers, whose task is to move packets closer to their destination. A router does not usually know the entire path to the far end of the journey. It simply reads the destination address and decides which of the neighbouring connections offers a good next step, then forwards the packet along that link. In this way a packet is handed on from router to router, each one bringing it a little nearer, until it finally arrives.

One of the most powerful features of this system is that the packets of a single message need not all follow the same route. If one path becomes crowded or a connection fails, routers can send later packets by a different way. The pieces of one email might travel across several different countries by several different cables, yet still arrive at the same place. Because no single route is essential, the network as a whole is remarkably resilient: damage to one part simply causes traffic to flow around the gap. This flexibility was one of the founding goals of the network's design, and it is a large part of why the internet is so hard to disable.

Sending the packets is only half the story. When they arrive, they must be checked and reassembled, and this is handled by a set of agreed rules known as a protocol. Two protocols work together so closely that they are usually named as a pair. One of them is responsible for breaking the message into packets at the start and putting them back in order at the end, while also making sure that nothing has been lost. If a packet fails to arrive, the receiving computer can notice the gap and ask for that piece to be sent again. This checking is what allows a large file to arrive complete and uncorrupted, even though it crossed the world in hundreds of separate fragments by many different paths.

Because people find long strings of numbers hard to remember, the internet also provides a system for translating names into addresses. When you type the name of a website, your computer first consults a service that acts like a giant directory, looking up the numerical address that corresponds to that name. Only once the address has been found can the packets be directed to the right place. This translation happens quietly in the background, usually in a fraction of a second, and most users never notice it at all.

What makes the whole arrangement so effective is that no single machine is in charge. There is no central office through which every message must pass and which, if it failed, would halt the entire system. Instead, control is spread across countless routers and computers, each following the same simple rules about how to read an address and where to pass a packet next. Out of these modest local decisions, repeated billions of times a second, emerges a global network capable of carrying almost any kind of information between almost any two points on Earth. The genius of the design lies not in any one clever component but in the way many simple parts cooperate.