Internet 3
Internet 3
Early networks
See all videos for this article
The first computer networks were dedicated special-purpose systems such as SABRE (an airline
reservation system) and AUTODIN I (a defense command-and-control system), both designed
and implemented in the late 1950s and early 1960s. By the early 1960s computer manufacturers
had begun to use semiconductor technology in commercial products, and both conventional
batch-processing and time-sharing systems were in place in many large, technologically
advanced companies. Time-sharing systems allowed a computer’s resources to be shared in rapid
succession with multiple users, cycling through the queue of users so quickly that the computer
appeared dedicated to each user’s tasks despite the existence of many others accessing the
system “simultaneously.” This led to the notion of sharing computer resources (called host
computers or simply hosts) over an entire network. Host-to-host interactions were envisioned,
along with access to specialized resources (such as supercomputers and mass storage systems)
and interactive access by remote users to the computational powers of time-sharing systems
located elsewhere. These ideas were first realized in ARPANET, which established the first host-
to-host network connection on October 29, 1969. It was created by the Advanced Research
Projects Agency (ARPA) of the U.S. Department of Defense. ARPANET was one of the first
general-purpose computer networks. It connected time-sharing computers at government-
supported research sites, principally universities in the United States, and it soon became a
critical piece of infrastructure for the computer science research community in the United States.
Tools and applications—such as the simple mail transfer protocol (SMTP, commonly referred to
as e-mail), for sending short messages, and the file transfer protocol (FTP), for longer
transmissions—quickly emerged. In order to achieve cost-effective interactive communications
between computers, which typically communicate in short bursts of data, ARPANET employed
the new technology of packet switching. Packet switching takes large messages (or chunks of
computer data) and breaks them into smaller, manageable pieces (known as packets) that can
travel independently over any available circuit to the target destination, where the pieces are
reassembled. Thus, unlike traditional voice communications, packet switching does not require a
single dedicated circuit between each pair of users.
Commercial packet networks were introduced in the 1970s, but these were designed principally
to provide efficient access to remote computers by dedicated terminals. Briefly, they replaced
long-distance modem connections by less-expensive “virtual” circuits over packet networks. In
the United States, Telenet and Tymnet were two such packet networks. Neither supported host-
to-host communications; in the 1970s this was still the province of the research networks, and it
would remain so for many years.