ARPANET

ARPANET is the foundation of the internet. Initiated in 1966, the Advanced Research Projects Agency Network (ARPANET) was the first wide-area network to employ packet switching. Incubated in the vital milieu of American's postwar cybernetic paradigm, ARPANET developed out of efforts to meet the organizational demands consistent with a burgeoning military industrial state facing a nuclear abyss.

Prehistory

Packet switching was developed by Paul Baran in the early 1960s under a RAND Corporation research project aimed at creating a communications network capable of surviving nuclear war. In Baran's own words:

If the strategic weapons command and control systems could be more survivable, then the country’s retaliatory capability could better allow it to withstand an attack and still function; a more stable position. But this was not a wholly feasible concept, because long-distance communications networks at that time were extremely vulnerable and not able to survive attack. That was the issue.

The problem is that conventional long-distance communications networks, such as the telephone network, operated in hierarchal and concentrated fashion. Calls are directed to a local, and, if necessary, a regional office, where they are routed manually. With each customer connected to only a single office, eliminating a single office would cut off a large number of nodes. So long as routing capabilities were centralized in offices, networks remained vulnerable. The Department of Defense, well aware of this, had already experimented with AUTOVON, designed and operated by the American Telephone and Telegraph Company (AT&T) between 1961 and 1963. AUTOVON did not represent a technological advance. It was a military voice network built atop the existing civilian telephone network. A sort of decentralization was achieved by increasing the frequency of switching centers; overall survivability was increased by fortifying them. By automating the routing process and distributing this capability to every node in the network, Baran's system eliminated the need for switch centers altogether. “The intelligence required to switch signals to surviving links is at the link nodes and not at one or a few centralized switching centers."

Packet switching made decentralized computer-routing possible by creating a system in which an address was attached to the communique itself, allowing each node to automatically and independently route any data they received. Packets consist of header and payload. Data in the header directs the payload to its destination, where the payload is processed by a separate operating system. Data flow is now a function of properties immanent to the message itself without the transcendental intervention of any operator outside the system. ARPANET is the real body without organs. Moreover, packet switching vastly improved the efficiency of communication systems.

Essential to early computing is the development of time-sharing. Issuing commands to primitive computers was a laborious process involving punch cards and frequent disappointment. Countless hours spent punching holes in cards and fiddling with magnetic tape- and all the while the computers sit idle.

Instead of running a single program from start to finish before going on to the next one, a time-sharing operating system would cycle between a number of programs, devoting a fraction of a second of processing time to each one before going on the next ... When a computer serves a user at a interactive terminal, it spends most of its time waiting for commands; very little time is spent actually processing data. If a computer can serve many terminals at once, it will spend less time idle and more time doing productive work, which increases the efficiency therefore the economically feasibility—of interactive computing.

By the mid 1960s, researchers regularly had access to time-sharing computers, and began to wonder if the principle behind time-sharing (high-speed cycling to simulate simultaneity) could be applied to communication between computers. Network communication suffered a temporal diseconomy similar to that of batch coding. Because computer messages typically came in short bursts separated by long pauses, computer users paid dearly for telephone connections that were inactive much of the time. Baran's packets- because they could be broken down into fractions, scattered, and reassembled independently- made possible the cycling operation necessary to efficiently distribute bandwidth.

Development and Implementation

Inspiration for the network was predictably bound up with postwar cybernetics. Joseph C. R. Licklider, veteran of SAGE and colleague of Norbert Weiner (father of cybernetics), was an early visionary. In 1960, he published a paper entitled "Man-Machine Symbiosis," calling for the implementation of thinking-machines and envisioning an “Intergalactic Computer Network”- an opened ended system designed for the transmission of data across vast distances. In 1962, Licklider was involved in the creation of a functional time-sharing system, working with Bolt, Beranek and Newman out of MIT. The system went into operation September 1962. In October of that same year, he was appointed head of ARPA's newly minted Information Processing Techniques Office (IPTO). The IPTO began as the Command and Control Research initiative, assigned to ARPA in June of 1961. According to internal documentation, the aim of the initiative was to "support research on the conceptual aspects of command and control and to provide a better understanding of organizational, informational, and man-machine relationships." In Licklider's view, the future of command and control would rely on advancements in computer science. In his own words:

There was a belief in the heads of a number of people -- a small number -- that people could really become very much more effective in their thinking and decision-making, if they had the support of a computer system, good displays and so forth, good data bases, computation at your command. It was kind of an image that we were working toward the realization of.... It really wasn't a command and control research program. It was an interactive computing program. And my belief was, and still is, you can't really do command and control outside the framework of such a thing... of course, that wasn't believed by people in the command control field.

Licklider's remained head of the program until 1964, and during this time the program underwent fundamental change. According to an ARPA sponsored internal history:

Emphasis had changed from command operational studies, war game scenarios and a "command systems laboratory" to research in time-sharing systems, computer graphics, improved computer languages, and computer networking... By the beginning of 1964 this change was reflected in renaming the office Information Processing Techniques.

Licklider left the IPTO in 1964, but his vision was carried on by Bob Taylor, who acquired funding for a network project in February of 1966. A plan was drafted and presented in October 1967 at the inaugural Symposium on Operating Systems Principles. It was at this conference that packet switching first came to the attention of ARPA investigators. Roberts reached out to both Baran and Donald Davies, an English researcher who had independently developed packet switching a number of years after the fact.

The network became operational in 1970. By 1975, operational control of the ARPANET passed to the Defense Communications Agency. In the same year, the NSA approved deployment of the first ARPANET encryption devices in order to support classified traffic. Research into packet encryption began during the early 1970s, carried out by Bolt, Beranek and Newman under DARPA funding.

On the occasion of ARPANET's decommissioning- February 28, 1990- Vinton Cerf, a key contributor to the project, penned "Requiem of the ARPANET:"

It was the first, and being first, was best,

but now we lay it down to ever rest.

Now pause with me a moment, shed some tears.

For auld lang syne, for love, for years and years

of faithful service, duty done, I weep.

Lay down thy packet, now, O friend, and sleep.