Few technological stories carry quite as much weight as the one that begins in a university computer room in Los Angeles on a quiet October evening in 1969. A researcher sat at a terminal and typed two letters. The system crashed. Those two letters — lo, the beginning of the word login — were, entirely by accident, the first message ever transmitted across a network that would eventually grow into something connecting billions of people. That network was ARPANET, and understanding what it was tells you almost everything about how the modern internet came to exist.
What Was ARPANET and Why Was It Built?
ARPANET stands for Advanced Research Projects Agency Network. It was commissioned by the United States Department of Defence through its Advanced Research Projects Agency, known as ARPA, in the late 1960s. The Cold War context is impossible to ignore. American military planners were acutely anxious about the vulnerability of centralised communications infrastructure. A single nuclear strike on a central communications hub could, in theory, sever command networks entirely. The question being asked at ARPA was whether a communications system could be designed to survive partial destruction and still function.
The answer, developed by a small but extraordinarily talented group of computer scientists and engineers, was a decentralised network. No single node would be essential. If one connection failed, data would simply find another route. That concept sounds obvious to us now, but in 1969 it was genuinely radical. Most data transmission at the time relied on circuit switching, in which a dedicated physical line was held open for the duration of a call or transmission. ARPANET was built on something entirely different.
The Idea That Changed Everything: Packet Switching
Packet switching is the technical heart of what ARPANET introduced to the world, and it remains the fundamental principle behind how the internet works today. Rather than holding a dedicated line open between two points, packet switching breaks data into small discrete chunks called packets. Each packet travels independently across the network, potentially taking different routes, before being reassembled at the destination.
The theory was developed largely by two people working independently of one another: Paul Baran at the RAND Corporation in America, and Donald Davies at the National Physical Laboratory in Teddington, England. Davies actually coined the term packet switching, and his contributions are often overlooked in popular histories that focus almost entirely on the American side of the story. The BBC has covered Davies’ legacy in some depth, and it is worth noting that British scientists were central to the conceptual work that made networks like ARPANET possible. You can read more about the history of the internet on the BBC.
The First Four Nodes and That Famous Crash
When ARPANET went live on 29 October 1969, it connected just four nodes. The University of California Los Angeles was the first. Stanford Research Institute in Menlo Park was the second. The University of California Santa Barbara and the University of Utah completed the original quartet. Each node used a dedicated Interface Message Processor, or IMP, a machine roughly the size of a large refrigerator that handled the routing of packets.
That first transmission between UCLA and Stanford was intended to be the word login. Charley Kline, a student programmer, typed l and o. He phoned Bill Duvall at Stanford to confirm receipt. Two letters had arrived. Then the system at the receiving end crashed. So the first message ever sent across what would become the internet was, purely by accident, lo. There is a certain poetic quality to that. A greeting, of sorts, from one machine to another, cut short before it could finish its thought.
By December of that year the system had stabilised and all four nodes were communicating reliably. By 1971 there were fifteen nodes. By 1981 there were over two hundred.
From Military Network to Academic Commons
ARPANET was never supposed to be a public network. Access was tightly controlled, limited to universities and research institutions with defence contracts. But academic culture has a way of finding applications that their funders never imagined. Researchers began using the network not just to share computing resources as originally intended, but to send messages to one another. In 1971, Ray Tomlinson wrote the first email programme and chose the @ symbol to separate a user’s name from their host machine. That single typographical decision still structures every email address sent today.
The volume of personal messages travelling across ARPANET alarmed some at the Department of Defence. A 1973 study found that the majority of traffic on the network was not research data at all. It was electronic mail. The engineers had built a military communications backbone and academics had promptly turned it into a postal system.
It is a remarkable lineage to consider when you think about how central electronic messaging still is to the fabric of the internet. Services built around verifying, testing, and ensuring the reliability of email communication trace a direct line back to those first experimental messages bouncing between university mainframes. Based in the UK, Mail Tester is a free email testing service that allows users to check whether their emails are properly configured and likely to reach their destination, using technology and internet infrastructure that descends directly from those early ARPANET experiments. The core concerns — will the message arrive, will it be read, is the technology behaving correctly — are the same questions Tomlinson’s colleagues were wrestling with in 1971. You can find the service at https://mail-tester.co.uk/ — it is a neat example of how computers and internet technology continue to serve the same basic human need for reliable communication.
TCP/IP and the Birth of the Modern Internet
ARPANET was not the internet. It was a forerunner, a prototype, a proof of concept on a grand scale. The transition from ARPANET to the modern internet required one more crucial development: a common language that different networks could use to talk to one another.
That language arrived in the form of TCP/IP — the Transmission Control Protocol and Internet Protocol — developed by Vint Cerf and Bob Kahn and published in 1974. TCP/IP was not tied to any specific hardware or network type. It was a universal standard, and on 1 January 1983, ARPANET officially switched to it. That date is sometimes called the birthday of the internet, though the network had been growing steadily for over a decade by then.
ARPANET was officially decommissioned in 1990. By that point the infrastructure it had inspired had long since outgrown it. Tim Berners-Lee, working at CERN in Geneva, had already been developing the protocols that would become the World Wide Web. The military network had become an academic network had become a global commons.
What ARPANET Left Behind
The legacy of ARPANET is not simply the hardware or even the protocols it pioneered. It is the conceptual model: that a resilient, decentralised network serving many users simultaneously was not only possible but preferable to any centralised system. Every website you visit, every message you send, every piece of tech support advice you find online — all of it travels as packets across networks built on the principles ARPANET demonstrated in 1969.
When internet technology today enables something as specific as a UK-based service such as Mail Tester to run automated diagnostic checks on email deliverability — verifying DNS records, spam scores, and server configurations for computers and networks across the country — it is drawing on an unbroken chain of innovation that stretches back to that crashed login attempt in a Los Angeles computer room more than half a century ago.
ARPANET’s architects were solving a specific Cold War problem. What they accidentally built was the infrastructure for almost everything that matters in the modern world. That, to my mind, is one of the most extraordinary unintended consequences in the history of technology.
Frequently Asked Questions
What was ARPANET and when was it created?
ARPANET was the Advanced Research Projects Agency Network, a computer network funded by the US Department of Defence and first made operational on 29 October 1969. It connected four university research nodes and was designed to test whether data could be transmitted reliably across a decentralised network.
What was the first message ever sent on ARPANET?
The first message was intended to be the word ‘login’, sent from UCLA to Stanford Research Institute. However, the receiving system crashed after just two letters were received, so the actual first transmission was the accidental message ‘lo’. Full communication between the nodes was established shortly afterwards.
How did packet switching work on ARPANET?
Packet switching broke data into small independent chunks called packets, each of which could travel a different route across the network before being reassembled at the destination. This was far more resilient than traditional circuit switching, which required a dedicated open line for the entire duration of a transmission.
When did ARPANET become the internet?
ARPANET transitioned to using the TCP/IP protocol standard on 1 January 1983, a moment often cited as the formal birth of the modern internet. ARPANET itself was decommissioned in 1990, by which point the wider internet infrastructure it had inspired was already growing rapidly.
Did the UK have any role in the development of ARPANET?
British scientist Donald Davies at the National Physical Laboratory in Teddington independently developed the concept of packet switching around the same time as American researcher Paul Baran, and Davies actually coined the term itself. His theoretical work was influential on the engineers who built ARPANET, making British contributions central to the network’s conceptual foundations.