Lexikon's History of Computing

ARPA and the ARPANET: A Brief History
ARPANET was the precursor of what we know of as the Internet today.

 Return to Title Page

ARPA - The Advanced Research Projects Agency:

The Financial Backbone of U.S. Computer Research

The single most influential agency in the history of computer development in the United States is the Advanced Research Projects Agency (ARPA). ARPA is the central research and development organization for the U.S. Department of Defense. Established in February 1958 by President Eisenhower, and later supported by the Kennedy Administration, ARPA's creation was in direct response to the launching of "Sputnik" by the former U.S.S.R.

The sobering realization that the Soviet Union had developed the capacity to successfully launch orbiting earth satellites and thereby exploit the regions of space for scientific and military purposes shocked the U.S. defense community and held the potential for escalated cold war tensions.

President Eisenhower responded to the growing need for a high-level organization within the Department of Defense to formulate and execute research and discovery projects to expand the frontiers of technology beyond the immediate and specific requirements of the military services and their laboratories.

ARPA's mission was to develop imaginative, innovative, and often high risk research ideas offering significant technological impact that went well beyond the normal evolutionary developmental approaches, and to pursue these ideas through to demonstrations of technical feasibility and development of prototype systems.

In 1960, all of ARPA's civilian space programs were transferred to the National Aeronautics and Space Administration (NASA) and the military space programs were transferred to their respective service (Air Force, Navy, etc.) This was a way of focusing the U.S. Space Program on specific objectives, while keeping ARPA focused in its main charter, which was funding research.

The IPTO and Computer Research Funding

ARPA became the largest funding agency for technical research and development. Within ARPA, a special office was established to support research dealing with the field of computers and computer related technologies. This was the Information Processing Techniques Office, or "IPTO."

In 1962, J.C.R. Licklider, a professor of Psychology at MIT, was hired by ARPA to become first Director of the IPTO. Licklider (known as "Lick") was a brilliant visionary and pioneer in the field of human-computer interaction and specifically the field of interactive computing. Licklider's work also laid the foundation for graduate work in the newly created field of computer science. (Licklider also worked on the SAGE computer project.)

Licklider pursued his vision of increased human-computer interaction and interactive computing despite criticism from some areas of the computer establishment. His perseverance led to breakthroughs in interactive computing and set the trend for further developments in this area.

Licklider's responsibilities at ARPA included selecting and funding researchers to build and lead research groups. By 1962, the value and potential of computers was becoming more widely recognized. However, computers were still extremely large, expensive and difficult to use. Part of Licklider's vision was a concept he originated called "man-computer symbiosis."

In a very influential paper written in 1960, Licklider outlined his vision of what the relationship of people to computers would some day become.

Licklider reasoned that to make computers more valuable, they also had to become more accessible and responsive to humans. One of Licklider's objectives included the creation and funding of several research projects exploring time sharing. The concept of time sharing would allow multiple individuals to share the high speed central processing power of a computer, thereby expanding the number of "concurrent" users the computer could serve. This not only increased computer availability to those who would otherwise have to wait in line for valuable computer time, but raised the level of importance of the field of human-computer interaction. As more and more people had access to computers, the way in which they communicated with computers became increasingly important.

Licklider was the architect of an early time sharing project called "Project MAC" (Machine Aided Cognition) at the Massachusetts Institute of Technology. Project MAC was oriented towards achieving a new level of human-computer interaction. The MAC program objectives included the development of flexible and responsive time sharing, closer machine-human interaction, better use of computers for education and research, and long-range development of national man-power assets through education. By 1963, Project MAC had implemented time sharing on an IBM 7094 computer, allowing it to accommodate 24 simultaneous users.

Around this time, Bell Telephone Laboratories (BTL) and General Electric (GE) joined with Project MAC in the development of MULTICS (Multiplexed Information and Computing Service). Bell Labs subsequently dropped out. GE developed the product for market. (Honeywell later bought GE's computing business). Honeywell provided MIT with a model 635 and a model 6080 computer, and Digital Equipment Corporation provided Project MAC with a DEC PDP-6.

Through ARPA and IPTO, Licklider also funded time-sharing projects at U.C. Berkeley ("Project Genie") and one at the Systems Development Corporation in Santa Monica. ARPA's annual funding for all three was approximately $15 million, which was a significant amount at that time.

Later on, research into time-sharing was also done by others including the RAND Corporation, Lincoln Laboratories, United Kingdom's National Physics Laboratory, and other locations.

At IPTO, Licklider also established a committee of various people and organizations who were involved in funding computer research in one form or another. Included on this committee were the Office of Naval Research, the National Institutes of Health, the Army, Navy, the National Aeronautics and Space Administration (NASA), the National Science Foundation (NSF) and others. The purpose of this group was, as Licklider saw it, to coordinate information on research projects and avoid unnecessary overlap.

Licklider's overall contributions to the field of computer science were significant. Prior to 1965, there were no U.S. universities granting a Ph.D. in computer science. Licklider, and his ARPA successors, provided funding for the research needed to create university graduate programs in computer science at U.C. Berkeley, CMU, MIT and Stanford. It wasn't until 1969 that Ph.D's in computer science were awarded.

In 1963, Licklider left IPTO and ultimately returned to MIT. He was succeeded as director of IPTO by Ivan Sutherland. Sutherland, with Licklider's encouragement, recruited Robert Taylor, of NASA, to become Associate Director of IPTO. Through NASA, Robert Taylor had been funding research work being done by Douglas Engelbart at Stanford Research Institute.

Dr. Engelbart was another visionary and pioneer in the field of human-computer interaction. Dr. Engelbart viewed computers as a way to augment human intellect, not just function as automation tools. With funding authorized by Robert Taylor, Engelbart established the Augmentation Research Center (ARC) from which grew some developments in the area of human-computer interaction. Most notable of these was the creation of the "mouse" interactive pointing device, the use of multiple windows displays and the focus on personal computing as a way to expand man's abilities.

Ivan Sutherland and Robert Taylor worked as a team to continue and expand upon the focus established by Licklider. They continued to push for funding of high tech research, and saw the IPTO's budget double in just a few years. The IPTO continued to provide funding for time sharing projects as well as computer graphics research, computer language research and early artificial intelligence (AI) projects. Those involved in some of the early projects included Marvin Minsky, John McCarthy, Jerry Feldman, Charlie Rosen, Nils Nilson and many others.

The IPTO focused on funding large projects and team based projects, as opposed to funding individuals. Typical project funding was in the range of $500,000 to several million per year. From 1965 to 1970, during the height of the Viet Nam war, overall ARPA funding was cut in half, while IPTO funding was doubled, which emphasized the importance that this type of funding was being given. By 1970, the total ARPA budget was $238 million, of which $26 million was dedicated to information processing research.

ARPA's Name Changes

In 1972, ARPA changed its name to DARPA for "Defense Advanced Research Projects Agency." In April 1993, it changed back to ARPA, dropping the word "defense" from its official title.

ARPA: 1976-1985

During the late 70's and early 1980's, ARPA's research focus included Command, Control and Communications (CIC) systems, tactical armor and anti-armor programs, infrared sensing for space-based surveillance, high-energy laser technology for space-based missile defense, anti- submarine warfare, advanced cruise missiles, advanced aircraft, defense applications of advanced computing, stealth technology, very large scale integration (VSLI), charged particle beam research and other areas. ARPA's research focus also included information processing, aircraft related programs, hypersonic research, satellites and submarine technology. In 1985, about $600 million was approved for strategic computer research in the U.S., including research into gallium arsenide semiconductors and parallel processing. ARPA's current focus is on revolutionary new technologies, including electronics and materials processing, computers, sensors, communication devices.

ARPA Today

In support of its mission, ARPA has developed and transferred technology programs encompassing a wide range of scientific disciplines which address the full spectrum of national security needs. ARPA manages and directs selected basic and applied research and development projects, and pursues research and technology where risk and payoff are both very high. The expert system DART, was used to assist the U.S. military in the logistics planning surrounding the Desert Storm operation. Today, ARPA is the lead Department of Defense agency for advanced technology research and has leadership responsibility for the HPCC (High Performance Computing and Communications) Program. The HPCC Program is designed to stimulate the aggressive advance of the high performance computing and communications technology base. The technologies include scalable technologies for high performance computing and networking along with system software to support the effective application of the high performance technologies to large scale problems.

ARPA continues its research into a variety of areas, including strategic computing for military applications, development of high performance networking technologies for the National Research Education Network (NREN), support to universities in computer science, high performance algorithms. In April 1993, its name changes back to ARPA.

(Information adapted in part from the Office of Science and Technology Policy (OSTP), and ARPA, Washington, D.C. 20506.)

Genesis of the ARPANET

In 1965, Ivan Sutherland left the IPTO and Robert Taylor became IPTO Director. As Director, Taylor had wide latitude in the area of project funding. Taylor was a strong advocate for expanding the use of computers, not just as number crunchers, but as a medium for people to externalize their ideas, observe them and communicate them. Taylor was convinced that computer communication would be the wave of the future. Taylor was well aware of the work being done by the various time-sharing projects in different parts of the country. He was also aware of the fact that these time sharing groups, or "communities," were not in sufficient electronic communication from group to group.

Time sharing interaction was limited to communication within a particular group, on a particular system. Access to another group or system required a separate computer terminal. At one time, Taylor had three different computer terminals in his office, to access three different time sharing systems. He realized that the real future of computer interaction and communication would require a network of computer systems that spanned systems and user communities.

In February, 1966, Taylor approached the then head of ARPA, Charles Herzfeld, and made a proposal for the creation of an interactive computer network, linking different computer systems together for the sharing of information. After a twenty-minute discussion with Taylor describing this proposal, Herzfeld approved of Taylor's concept and provided $1 million in initial project funds. Taylor's concept of computer-to- computer communication grew in time to become the largest computer network in the world-- ARPANET. In late 1966, Taylor had decided upon Lawrence Roberts, then working at Lincoln Labs, as his choice for a program manager of the ARPA project. Lawrence was reluctant to join what he felt was a bureaucratic organiza- tion, and initially declined Taylor's job offer. However, Taylor realized that ARPA funded over half of the work being done by Lincoln Laboratories. Taylor informed Charles Herzfeld, head of ARPA, of his desire to hire Lawrence Roberts. He suggested that Herzfeld call the head of Lincoln Labs and stress that it would be in Robert's best interest to take the job at IPTO offered by Taylor. Herzfeld made such call, and the reluctant Roberts was added to the IPTO team in December 1966.

The consulting and development firm of Bolt, Beranek and Newman (BB&N) won the contract to provide the interface message processors (IMP). Other work on ARPANET was also done by DECCO, and Network Analysis Corporation. Taylor saw the need to get private sector companies involved in time sharing projects and activities and emphasized the need for technology transfer among the scientific community. AT&T had been approached with an invitation to participate but thought that packet-switching technology would never work. In 1967, Taylor organized a large European computer conference, hosted by the National Physics Laboratory in England, to introduce time sharing to Europe and encouraged and funded some of those involved in ARPA-funded projects to travel to the conference and share their experiences. The ARPANET had its first three nodes in operation by 1969. Development of ARPANET involved individuals and institutions in the computer research communities which were supported by the growing ARPA IPTO program. Among the related early work done by others was the OCTOPUS system, developed by Lawrence Livermore Laboratories, which linked large computers. Bell Telephone Laboratories was conducting experiments in linking computers, the SITA airline reservation system was being developed, and Dartmouth College had its NERCOMP system, which linked a number of smaller academic institutions throughout New England.

The Evolution of ARPANET

From the early 1970's, ARPANET packet-switching technology has been the basis for the development of defense-wide systems for data communications. Public availability of technical information on the development and operation of ARPANET led to progress by others in this same field.

In the period between 1969 to 1975, there was a great increase in the number of computer networks in use. In the U.S., the Federal Communications Commission (FCC) decided to permit value-added carriers to compete with established carriers, thus improving the economics of networking. By 1990, ARPANET formed a major part of the worldwide network called INTERNET.

INTERNET is a collection of interconnected networks, linking universities, government laboratories, military bases and commercial organizations. By 1990, Internet had over 1,000 active networks with over 100,000 connected computers.

Today, the Internet is estimated to have well over 50 million users.

See Also:

Individuals: Computer Pioneers, Scientists, Entrepreneurs

Interesting Computer Facts

Chronology

Glossary

Table of Contents

Master Index

 

 

Return to Title Page

Copyright © 1982-2003, Lexikon Services "History of Computing" ISBN 0-944601-78-2