In the 1970s, scientists and engineers imagined a world where information could be exchanged worldwide in the blink of an eye—and used their ingenuity to make it a reality.
The 21st-century world now extends into online space, a rich environment of communication, information and entertainment that anyone can access with a computer, tablet or phone. But to reach that world we had to solve a problem—how to get computers to communicate directly with each other.
The first networks connected machines in the same building, then the same city, then the same country. In 1973, for the first time, a data network spanned the Atlantic Ocean. Within a few years, engineers designed software that allowed different networks to communicate with one another, and the internet was born.
Computers come in from the cold
Driven by the Cold War, the US and USSR had built rival supercomputers, the biggest and fastest calculators in the world. Designed to simulate explosions, crack codes and consolidate surveillance data, they were difficult to use, unconnected to other machines, vulnerable to attack and accessible to only a few people.
It made sense to build them into a network so that the power of their processing could be shared. From 1957 IBM built the Semi-Automatic Ground Environment (SAGE), a system designed to coordinate defence against Soviet attack from the air.
Working on the interface between machines and their human operators on this project was a computer engineer and psychologist from the Massachusetts Institute of Technology (MIT), J C R Licklider. In the early 1960s he moved to head the computer division at the US government's Advanced Research Projects Agency (ARPA), set up in the wake of the launch of the Soviet satellite Sputnik.
Beyond the immediate military applications, he believed that the role of computers in society could be transformed if they were combined into networks, and encouraged his colleagues to take the idea forward.
Expanding the network
The result was ARPANET, the forerunner of the internet, initiated in 1967 and operational two years later. In these early days it was largely seen as a tool for academic engineers and computer scientists, linking departments at the University of California at Los Angeles, the Stanford Research Institute, the University of California, Santa Barbara, and the University of Utah.
By 1972 the network had grown to include over 30 institutions, encompassing government offices and businesses as well as universities, and stretching from the Pacific coast to the Atlantic. As well as sharing data, the network also made it easier for scientists to communicate with each other through email, which became one of its most popular applications.
A transatlantic data exchange
The success of the network depended on a new technology called 'packet switching', the idea behind which originated independently on either side of the Atlantic.
Donald Davies, manager of the Advanced Computer Techniques Project at the UK's National Physical Laboratory (NPL), proposed a network that would enable businesses to exchange data across the country. He coined the term 'packet switching' to describe a method of transmitting the data efficiently by breaking it into smaller units, which would travel independently by the next available route before being recombined to pass on to their destinations.
At the same time Paul Baran, working at the Rand Corporation in the US, designed a communications system for the US military that transmitted data via a network with multiple nodes. The idea was that if enemy attackers knocked out one or more nodes, the system as a whole would have the resilience to continue transmitting. He also suggested splitting the data in the smaller units, which he called 'message blocks'.
On 25 July 1973, a computer at University College London (UCL) sent 'packets' of digital information to another in California. Led by Peter Kirstein at the Institute for Computer Science, UCL, this was ARPANET's first transatlantic link. It ran from London to Norway via cable, then on to the Seismic Data Analysis Centre (SDAC) in Virginia via satellite link, and finally across the US network to the Information Sciences Institute at the University of Southern California.
The ARPANET team had originally identified NPL as its first London link, but it was deemed politically unacceptable, in the context of developing relationships within Europe, for a British government research laboratory to connect directly with the US Department of Defense.
The universal language of computers
A problem arose very quickly as other organisations set up their own networks, which were incompatible with ARPANET and each other. What was needed was a universal language that allowed information to pass between networks, and indeed between any two networked computers, regardless of their hardware and software.
The answer came with the Transmission Control Protocol (TCP) and the Inter-network Protocol (IP) developed in the mid-1970s by Bob Kahn and Vint Cerf for ARPA. Known collectively as TCP/IP, the protocols require every computer in a network to have a unique address, known as its IP address, and include services and applications to make interconnection possible.
Less than a decade after the first computers began operating, a single 'inter-network'—or internet—provided a universal highway for computers to communicate.
See all objects on display in Information Age: Web at the Science Museum