A Little Internet History…
Written by David Ferro's Social Implications of Computing course over the early 2000s
By the end of the 20th century computer “hosts” (on each of which might be many users) connected to the Internet numbered greater than 120 million. Its potentiality exploded into public consciousness after the release of the World Wide Web (1991 by Tim Berners-Lee at CERN research institute) and the first browsers (Gopher in 1991 and Mosaic in 1993); software programs that created a intuitive user interface with the hyperlink that allowed for viewing of information such as text and graphics on remote machines via the Internet. The origins of the Internet began three decades earlier, at the height of both the Cold War and negative reaction to the Vietnam War, as the culminated efforts of individuals responding to personal, technical, and societal challenges. Management of the initial project began at the Advanced Research Projects Agency (ARPA), a somewhat free wheeling academic agency that contrasted considerably in its organizational approaches with its more hierarchically organized Pentagon sponsor. In the end, this culture weighed heavily in the rapid and successful creation of the hardware, software, and protocols that connected the host machines and the Internet’s continual evolution.
President Dwight D. Eisenhower created ARPA in early 1958 as part of a response to the perceived threat of the Soviet’s launch of Sputnik in October of 1957. Buoyed by the efforts of scientists in World War II in creating the Atomic Bomb, the ENIAC machine, and breaking the German ENIGMA code, among other monumental inventions, Eisenhower felt that given free reign, American scientists would provide the flexible response necessary in dealing with the Soviets. In 1962, a scientist from MIT’s Lincoln Lab, J.C.R. Licklider (“Lick”) joined ARPA to take charge of the center’s computing research. He converted the office from command and control research, basically wargames, to the Information Processing Techniques Office (IPTO) devoted to useful computer research. In 1962 he writes a memo indicating that computers need to communicate despite all their different languages but the project never materialized. Bob Taylor held IPTO position in 1966, noted the three computer terminals each needed to connect to three different machines in different locals around the U.S. and quickly revitalized Lick’s idea, securing one million dollars in funding. Taylor’s argument: research universities working with IPTO were looking for more resources. Why not allow sharing of resources. Also, with the government buying different machines that vendors had no incentive to make communicate with one another, this might allow the government to circumvent that problem. Taylor also decided that they should build the system so that if a line went down, communication could route through different line and to start with 4 machines in different locals and build up to 16. Taylor recruited 29-year- old Larry Roberts to build it late in 1966.
In 1967, Roberts began the process of convincing a number of research universities receiving ARPA funding for other projects to join a resource-sharing network (dubbed ARPANET by one scientist – the name that stuck). The architecture began to take shape. The universities initially showed great reluctance at sharing their precious computing resources and that reluctance likely contributed to what became a key technology in connecting the machines. In order to lessen the load on the computers, Wes Clark of Washington University of St. Louis proposed the Interface Message Processor (IMP): a separate simpler computer for each computer on the net that would handle the network communication. Through a series of meetings and conferences, other critical ideas emerged, some from disparate areas. Paul Baran of the RAND corporation had been working on the vulnerability of the U.S. telephone communication system since 1960 but had yet to convince the telephone monopoly AT&T of his ideas. He had devised a scheme of breaking signals into blocks of information that would reassemble when reaching their destination. In addition, these blocks would travel through a “distributed network” where each “node” or communication point didn’t just connect to a centralized point but to 3 or 4 other nodes – as analogy, like a fish net rather than an astrisk. Each node could independently decide which path the block of information would take. Completely independently, in 1965, Donald Davies at the British National Physical Laboratory (NPL) came up with a similar concept he termed “packet-switching” with “packet” similar to Baran’s blocks. Unlike two people having a voice conversation on a devoted phone line, data typically comes in bursts, leaving a devoted line idle for much of the time. Davies didn’t like the inefficiency of the phone companies circuit-switching approach for data. In 1965, Tom Marill in a proof of concept project had connected two computers (one in Massachusetts and one in Santa Monica) using a Western Union wire and a “modem” on each end. Marill invented a “protocol” – a set of procedures where a computer would send a message to another computer and if it did not receive an acknowledgement of receipt would send another copy. These ideas and attendant terminology would form the basis of the ARPANET.
Indicative of ARPA’s shyness of bureaucracy, the smallest bidder for the IMP, Bolt Beranek and Newman (BBN), a small consulting firm in Cambridge, MA, got the contract to construct the IMP in December of 1968. They decided that the IMP would only handle the routing, not the data content of information transmitted (as analogy: the envelop but not the letter). In 1970, this left the development of how the computers would understand each other – the “host-to-host” protocols and software - up to the professors and graduate students at the host universities. They called themselves the Network Working Group (NWG) and began to share information via Request For Comment (RFC) papers. Steve Crocker at UCLA wrote RFC Number 1 titled “Host Software” in April 7, 1969. As the machines didn’t know how to talk “peer-to-peer” the researchers wrote programs that fooled the computers into thinking they were talking to dumb terminals. By summer they had agreed on an overall protocol called Network Control Protocol (NCP). RFCs are still used today to communicate issues of Internet protocol.
ARPANET began with the first installation of an IMP at UCLA quickly followed by Stanford Research Institute (SRI), University of California at Santa Barbara, and the University of Utah in the Fall of 1969. The first message transmitted between UCLA and SRI was “L”, “O”, “G”, the first three letters of “LOGIN” before the system crashed. The system worked remarkably well after that, however, and grew at a node a month in 1970. Inevitable improvements and similar networks followed. BBN invented remote diagnostics in 1970. In 1973, DARPA (ARPA renamed) divested itself of ARPANET. In the mid 1970s the National Science Foundation created CSNET and then in 1985 created the NSFNET “backbone” that all networks could connect to. NCP is replaced by Transmission Control Protocol / Internet Protocol (TCP/IP) in 1982 and that becomes a standard for all machines on all networks by mid to late 1990s. The open architecture of the what is now called the “Internet” with a capital “I” – a network of networks – allows for numerous innovations that propel it forward from contributors all over the world. The robust version of a means to move files – File Transfer Protocol (FTP) – is detailed in RFC 354 by Jon Postel. Attached to it is a means to send messages (later dubbed e-mail). Within a year, 2/3rds of traffic is determined to be messages. The resource sharing capability of the network becomes secondary. By 1989, ARPANET is dismantled but the Internet is thriving.
Some informative links:
www.isoc.org/internet/history
Many links, including some below (although, not Sterling’s, Taylor opposes it)
https://mel.lib.mi.us/internet/INET-history.html
Includes the following:
https://www.cc.ukans.edu/~grobe/early-lynx.html
The history of lynx, the first (text only) browser
https://www.w3.org/History.html
History of WWW, starting with V.Bush. (quite brief)
https://www.netvalley.com/intval.html
Starts with Sputnik spurring ARPA
https://www.thocp.net/main.htm
The history of computing
https://www.zakon.org/robert/internet/timeline/
Hobbes Internet timeline (Quick overview) starts with Sputnik