A network of networks
The internet is not a single network owned by one organization. It is a collection of independently run networks that agree to exchange data using common rules. That arrangement is the key to its history: the achievement was less about building one big network than about getting many different ones to interconnect. The web, which many people mistake for the internet, is a later service built on top and is covered in our piece on the early web.
The problem of sharing computers
In the early 1960s, computers were expensive and shared. Researchers wanted to connect them so that scientists in different places could use each other’s machines and data. Traditional telephone networks set up a dedicated circuit for each call, which was inefficient for the short bursts of traffic typical of computers.
The answer was packet switching. Instead of sending a continuous stream along a reserved line, a message is chopped into small packets that travel separately through the network and are reassembled at the destination. Paul Baran in the United States and Donald Davies in Britain developed the concept independently in the 1960s, and it gave networks flexibility and resilience, since packets could take different routes.
1969: ARPANET begins
The U.S. Defense Department’s Advanced Research Projects Agency funded a pioneering experimental network, ARPANET. Its first connections linked university and research sites in 1969, beginning with UCLA and the Stanford Research Institute. The first attempted message reportedly crashed partway through, but the link soon worked. By the end of 1969, there were four nodes. ARPANET’s popular reputation as a network built to survive a nuclear war is a simplification; its main purpose was to let researchers share resources, though resilient design ideas were certainly part of the wider thinking.
In 1971, Ray Tomlinson sent what is generally regarded as the first network email between machines, using the @ sign to separate the user name from the machine. Email soon became the most popular use of the network.
Many networks, one set of rules
By the early 1970s, other networks existed, including radio-based and satellite-based experiments, and they did not speak the same language. Vint Cerf and Robert Kahn described a method for interconnecting different networks in a 1974 paper, and their ideas developed into TCP/IP, a pair of protocols. IP handles addressing and forwarding of packets, and TCP provides reliable delivery on top. A design principle was to keep the network core simple and let the machines at the edges handle the complicated parts.
On January 1, 1983, ARPANET switched to TCP/IP, a date often called the internet’s birthday. That year or soon after, the domain name system was designed so that people could use names like example.com instead of numerical addresses. The same period saw the growth of university and scientific networks, and in the later 1980s, a national research backbone in the United States connected many institutions.
From research to the public
During the 1980s, internet use remained mostly academic and technical. Policy restrictions on commercial traffic eased in the early 1990s, and commercial providers began to offer access to businesses and individuals. Dial-up modems let households connect over telephone lines. When the web and graphical browsers appeared in the early 1990s, interest grew rapidly. Search tools, online shopping and email accounts attracted millions of new users within a few years.
The rapid growth raised practical issues. Address space started to run short, and a newer version of the addressing system, IPv6, was designed with a much larger capacity, though the transition has taken a long time. Traffic routing between networks relies on agreements and protocols that were originally built for trust among small communities, and security problems showed that the internet had not been built with attackers in mind. This led to the development of encryption basics for web traffic and of wider cybersecurity basics for ordinary users.
Broadband and mobile
In the 2000s, faster home connections replaced most dial-up service, using upgraded telephone lines, cable television systems or fiber-optic cable. High-speed access supported music and video streaming, and later cloud computing services. At the same time, wireless networks and smartphones made constant access normal, and the internet reached many places that never had wired computers.
Beneath all this lies physical infrastructure: undersea cables, long-distance fiber, exchange points and data centers. The physical layer is explained in our article on internet infrastructure. Satellite communication supplements it, especially in remote areas, though capacity and delay differ from fiber.
Governance and openness
No single entity runs the internet. Technical standards come from volunteer-driven bodies that publish open documents, and other organizations coordinate names and numbers. This openness let new services appear without permission, from messaging to video, but it also makes decisions slow and contested. Debates continue over privacy, access, regulation and who bears the cost of carrying traffic.
The lesson of the timeline
The internet’s history shows how a modest research project became a general-purpose layer for communication. Three ideas stand out: packet switching, which made networks efficient and flexible; shared protocols, which let separate networks interconnect; and open standards, which let anyone build on top. These choices, made decades ago, help explain why the network could grow from four nodes to a worldwide system.