The Cold War is often remembered for its fallout shelters, its spy thrillers, and the long shadow it cast across the history books. One of its most lasting legacies, however, is far less dramatic and far more present in daily life: the internet. The global network that now carries our emails, video calls and social media began as a response to one of the most unsettling questions of the nuclear age.
A Question of Survival
In the early 1960s, American defence planners faced a troubling problem. If a nuclear strike destroyed key parts of the nation's communication infrastructure, could the military and government still talk to one another? The telephone network of the time offered little reassurance. It relied on circuit switching, in which a dedicated, continuous line was established between two parties for the full length of a call. These circuits passed through centralised switching stations, and if one critical link or station was destroyed, the connection failed entirely. A small number of well-placed strikes could, in theory, leave commanders unable to communicate.
Researchers began to imagine a different kind of network, one with no single point of failure. At the RAND Corporation, engineer Paul Baran proposed a distributed design in which every node was connected to several others, so that the loss of any one would not isolate the rest. Around the same time, British computer scientist Donald Davies, working independently at the National Physical Laboratory, arrived at a similar idea and gave it the name by which it is still known: packet switching.
The Packet Revolution
The principle behind packet switching was both simple and radical. Instead of sending a message as one unbroken stream, the network would divide it into small blocks of data called packets. Each packet carried information about its destination and could travel independently, taking whatever route was available at that moment. On arrival, the packets would be reassembled into the original message.
The advantages were considerable. If one path was damaged or congested, packets could be redirected through another. Network capacity was also used far more efficiently, since many conversations could share the same lines rather than each monopolising a dedicated circuit.
ARPANET: The First Heartbeat
These ideas took practical form through the Advanced Research Projects Agency (ARPA), an agency of the U.S. Department of Defense. Although concerns about resilience shaped the thinking of the era, ARPA's immediate goal was also practical: to allow researchers at different institutions to share scarce and expensive computing resources.
In 1969, ARPANET connected its first four nodes: the University of California, Los Angeles; the Stanford Research Institute; the University of California, Santa Barbara; and the University of Utah. On 29 October of that year, a student at UCLA attempted to send the word "LOGIN" to Stanford. The system crashed after the first two letters, and so the first message ever transmitted across the network was, fittingly, "LO." Within hours the problem was fixed and the full login succeeded. The internet's first heartbeat had been recorded.
The Problem of a Common Language
ARPANET grew steadily through the early 1970s, but it had clear limitations. Its computers communicated using the Network Control Program (NCP), a protocol that worked only within ARPANET itself. Meanwhile, other networks were emerging, including packet radio and satellite networks, each with its own technical characteristics. NCP could not link these separate systems together, and it had no robust way of recovering when data was lost or arrived out of order.
Two researchers, Vint Cerf and Bob Kahn, set out to solve this problem. In a landmark 1974 paper, they described a method for connecting different networks into a single "network of networks," the concept from which the word internet derives. Their design eventually separated the work into two distinct layers:
Internet Protocol (IP) handled addressing and routing, ensuring that each packet carried the information needed to find its way across multiple networks to the correct destination.
Transmission Control Protocol (TCP) managed reliability, checking that all packets arrived intact and in the correct sequence, and requesting retransmission of anything lost along the way.
This division of labour proved remarkably powerful. Because IP made few assumptions about the underlying networks, almost any kind of network could be connected to it. Together, the two protocols became known as TCP/IP.
The Day the Internet Was Born
After years of development and testing, ARPANET formally adopted TCP/IP on 1 January 1983, a transition often called "flag day." Every connected computer was required to switch from NCP to the new system. Many historians regard this date as the true birth of the modern internet, since it marked the moment a common, open language was established for interconnected networks.
The design was so flexible and so robust that it has never been replaced. Over the following decades, networks were added at an extraordinary pace, from universities and research centres to businesses and, eventually, households worldwide. The invention of the World Wide Web by Tim Berners-Lee in 1989 built on top of this foundation, making the internet accessible to ordinary users.
A Quiet Legacy
More than forty years later, the principles conceived during the anxieties of the Cold War still operate beneath nearly every digital interaction. Each email, streamed film and video call is broken into packets, dispersed across countless routes, and reassembled at its destination, exactly as early researchers envisioned. A system designed with resilience in mind, capable of surviving the loss of any single part, has become the most widely used communication network in human history. In this sense, one of the most enduring products of an era defined by fear is a technology that now connects the world.