26 October 2019 | New Scientist | 35information. Funded by the US Department of
Defense, the UCLA and Stanford computers
were the first two nodes of this network. By
December 1969, two others had been installed:
at the University of California, Santa Barbara,
and the University of Utah in Salt Lake City.
In 1973, ARPANET went international,
connecting via satellite to nodes at the
Norwegian Seismic Array in Kjeller near
Oslo and University College London. Today,
a backbone of fibre-optic cables under sea
and land, supplemented by satellite links
and lower-tech copper telephone wires,
ensure near-global coverage (see “A global
network”, page 38).
Key features of how the modern internet
works were there right from these small
beginnings. Crucially, there was no centralised
control. ARPANET was a distributed “network
of networks”. Information, broken into
hundreds or thousands of small packets,
travelled from node to node through or
between these networks. If one node went
offline, the information would find another
way through, with each packet basing its
trajectory on feedback from previous ones.
This concept, known as packet switching,
had been developed in the early 1960s by
three independent groups of researchers in
the UK and US, including Kleinrock’s team.
“It made for a very resilient system,” says
Johnny Ryan at tech firm Brave, author of
A History of the Internet and the Digital Future.
“These packets are blindly going through the
network trying to find a quick route.”
Shared communication required a shared
language. That came in the form of a set of
standards known as TCP/IP – the Transmission
Control Protocol and Internet Protocol – first
made public by computer scientists Vint Cerf
and Bob Kahn in 1974 (see “The internet
evangelist”, page 42). These covered, among
other things, the standard format of data
packets and a unified system of addressing
so that networks could identify one another.
Such IP addresses are still assigned to all
networked computers today.
“That was a breakthrough,” says Wendy Hall,
a computer scientist at the University of
Southampton, UK. Open and free, TCP/IP
enabled anybody to put a computer on
the network, and any computer to talk to
another. On 1 January 1983, ARPANET adopted
it as its standard for “internetworking”,
and the modern internet was born. >LIAM MADDENI
T BEGAN – some would say, as it meant to
go on – with an error message. Late on the
evening of 29 October 1969, student
programmer Charles Kline attempted to send
some text from a computer at the University
of California, Los Angeles (UCLA), to another
at the Stanford Research Institute, more than
500 kilometres up the Californian coast.
“LOGIN”, it was supposed to say. Kline
got as far as “LO” before the system crashed.
The full message was resent an hour later.
What would eventually morph into the
largest communications network in human
history had made its debut: the internet.
It is fair to say that no one there quite
appreciated the full scope of what had
happened. “We knew we were creating an
important new technology that we expected
would be of use to a segment of the population,
but we had no idea how truly momentous
an event it was,” Leonard Kleinrock, Kline’s
supervisor, later said. Fifty years on, we are
still only just beginning to come to terms
with the consequences.
The Advanced Research Projects Agency
Network, or ARPANET, as the internet’s
precursor is better known, was an academic
project intended to allow computers to share