|It is one if the greater ironies of computer history that the vision of the hippies had to be married to the technology funded and fostered by the Pentagon, before the world of integrated information and communication became possible. [Stephen Segaller, NERDS2.0 p 134]|
J.C.R. Licklider's vision of man-computer symbiosis drove the work of the Information Processing Techniques Office of ARPA in the early 1960s. In memos addressed to The Members of the Intergalactic Computer Network, Licklider presented a futuristic vision where computers could be tied together in a common network using standard languages, enabling the computer scientists to share and build on each other’s work. In 1964, Robert Taylor took over for Licklider, and brought on Larry Roberts to build the network.
Unlike Baran, who was attempting to solve the communications problems of the entire Department of Defense, Roberts had a more modest goal: network together a few ARPA funded computer centers. Roberts was not so much building a network, as he was networking computer researchers. That is an important difference, as the important stuff was the ends of the network, not the network itself. The job of the network was to provide connectivity (and interconnection), support the work of the research and innovation at the ends, and otherwise get out of the way. In 1967, Roberts released a blueprint for the ARPANET entitled Multiple Computer Networks and Intercomputer Communications. ARPANET would be packet switched using shared telecommunication paths; distributed (resilient to failure); decentralized (no central gatekeeper); able to freely scale (assets could be easily added without permission); and operated with a sub-network of computers whose sole job was to route traffic (not process data).
The ARPANET was launched in 1969, and by 1972 the "experiment" was causing a stir. Other networks quickly followed. ARPA recognized the need to interconnect incompatible computer networks. Therefore, Vint Cerf and Robert Kahn developed the Internet Protocol, a "simple" protocol that created “a network of networks,” providing interconnection between otherwise incompatible computer resources, and facilitating research at the ends of the network. On January 1, 1983, ARPANET migrated to the Internet Protocol.
The Internet's design was mundane; it was designed to interconnect and do as little else as possible. It was a simple protocol in the middle designed to work over any physical communications network (telephone lines, radio, cable, fiber) and to support any application (e.g., file transfer, email, World Wide Web, video, VoIP). This created what was known as the “hourglass” design of the Internet. Incompatible networks and applications are networked together by the Internet in the middle. On the bottom of the hourglass is the diversity of different networks; on the top is the diversity of different applications. In the middle is the Internet Protocol. Applications need not be written for any particular network; as long as they are written to work over the Internet Protocol, they will work over any network. Networks do not need to be written to support or optimized for any particular application; as long as they support the Internet Protocol, any higher level application will work. Instead of having multiple network terminals in an office, each optimized to connect to a specific network that did specific things, an individual would have one terminal connected to one network that could support whatever the end user wanted to do. That network did not engage in processing of data, did not interact with the transmissions, and did not itself process any of the computations. The network just moved the bits, leaving the work for the muscle on the ends. [See also End-to-End Design; Layered Model]
Based on visions of what a computer network could be, ARPA made practical decisions on what the network needed to be. These design decisions set the foundation for the Internet. These decisions evolved from "what we need the network to do," to "network design principles" (1981), to "policy objectives" (1999), and regulatory requirements (2010).
Robert Taylor succeeded Sutherland as director director of IPTO in 1966. Taylor started at NASA in 1961. Robert Taylor's office in the Pentagon was Room 3D-200. [Roberts, Net Chronology]
Taylor receives approval from ARPA Director Charles Herzfeld and $1 m funding. [Waldrop 80]
While at Lincoln Labs, L. Roberts & T. Merrill, "Toward a Cooperative Network of Time-Shared Computers", Fall AFIPS Conf., Oct. 1966. Demonstrate an early computer network; funded by IPTO. [Roberts, Net Chronology] [Abbate p 48] [Heart 1990]
DOD funded MIT Lincoln Lab and System Development Corp in Santa Monica to demonstrate a cross country network. The linking of the computers was successful. However, Roberts stated "The problem... was with using telephone circuit switching technology because the throughput was inadequate to accomplish their goals." [Hauben] See ARPANET Telecommunications
Robert Taylor persuaded ("blackmailed") Larry Roberts moved from MIT to DARPA to be the Chief Scientist of the ARPANet. [Roberts, Net Chronology] [Abbate p. 44] [Cerf, Oral History 1990] [Roberts, Computer Science Museum 1988 ("when they first approached me, my response was: "No, go away. I'm having great fun doing research,"")]
IPTO funding 17 computer sites. [Salus p 19]
Frank Heart and Robert Kahn joins BBN [Vanity Fair] [Heart 1990]
Don Davies publicly presents his ideas on packet switching, is introduced to the work of Paul Baran. One of Davies design objectives was to build a communications network for data communications. A packet switched network would be a commercial service competing along side other communications networks, optimized for computer and data traffic, and targeting business consumers. Responsibility to build an experimental network was given to the UK General Post Office. [Abbate p 27-28] [Heart p 24 1990 (discussing working with Davies)]
FCC initiates the Computer Inquiry proceedings.
U.S. NAT'L COMM'N ON TECHNOLOGY, AUTOMATION AND ECONOMIC PROGRESS, TECHNOLOGY AND THE AMERICAN ECONOMY 252 (1966) (recommending "the federal government bring into being as rapidly as technology permits, at least one limited but major information processing network that is planned, developed, and operated to: (a) accelerate technological advance and gain experience in appraising economic and social benefits and costs of information networks; and (b) help meet a recognized unmet national need, such as for better information transfer through a national network of libraries and specialized information centers.")
D. W. Davies, K. A. Bartlett, R. A. Scantlebury, & P. T. Wilkinson, A Digital Communications Network for Computers Giving Rapid Response at Remote Terminals, ACM Symp. Operating Systems Problems, (Oct. 1967).
Manley R. Irwin, The Computer Utility: Competition or Regulation? 76 Yale L.J. 1299, 1318 (1967) ("There is a strong argument that a government data processing network would bring down computer costs and accelerate technical progress. But as a political proposition, the government network is a hopeless scheme.")
Roger Scantlebury deliver paper (written with Donald W Davies, at the National Physical Laboratory, England) "A digital communication network for computers giving rapid response at remote terminals." (first reference to "packet") (Abstract: "Those computer applications which involve rapid response to events at distant points create special problems in digital communication. Such applications are increasing in number, and could increase more rapidly if better solutions existed to the communication problems. The present-day methods for communication of data in rapid-response systems employ 'private wires' for the transmission paths or, where the available data rate and reliability is sufficient, employ voice channels from the switched telephone network. Given these rather arbitrary transmission facilities the user adds the terminal equipment necessary to make a communication system and sometimes integrates a number of paths into a private network."
April: IPTO Principal Investigator Meeting, ARPANet Design Session in Ann Arbor, MI. Wesley Clark recommends use of the Interface Message Processor (subnetwork of micro computers) [Clark] [Roberts, Net Chronology (" It was at this meeting that Wes Clark suggested the use of mini-computers for network packet switches instead of using the main frame computers themselves for switching.")] [Roberts, Computer Science Museum 1988 ("he said: "No, don't do it in the central computer, do it in the little computer."")] The design objective was to have the mainframes only interact with the IMPs, and have the IMPs do all the network work. The result was a subnetwork dedicated to networking functions without any processing at the application or content layers. [Salus p 19] [Waldrop 83] [Abbate p 49, 52 (discussing adoption of layering in ARPANet design]
June: Larry Roberts discusses preliminary plans of the ARPANET which would include 35 computers at 16 locations. Lawrence G. Roberts, Multiple Computer Networks and Intercomputer Communications (June 1967), Proceeding SOSP '67 Proceedings of the first ACM symposium on Operating System Principles Pages 3.1 - 3.6
The suggestion of networking mainframes for the purposes of sharing computer resources was not received with enthusiasm. Schools were hesitant to share their prized DARPA funded mainframe computers with other schools. However, since these computer resources were ARPA funded, ARPA could use this funding source in order to leverage and persuade the recalcitrant universities to participate in the ARPANet and play nice with the other participating schools, sharing their resources. [Babbage 23, 24] [Abbate p 49]
October: ACM Symposium on Operating Systems Principles Gatlinburg, Tennessee [Hauben]
Larry Robert's plan for ARPANet, Multiple Computer Networks and Intercomputer Communication [Roberts, Net Chronology]
Referred to network as "ARPANet" for the first time. [Segaller 62]
Larry Robert's plans relies upon the earlier work of Leonard Kleinrock. The transmission likes would be 2400 bps and ASCII was specified. The network would connect 16 research groups [Salus p 19] [Hauben]
Abstract: "There are many reasons for establishing a network which allows many computers to communicate with each other to interchange and execute programs or data. The definition of a network within this paper will always be that of a network between computers, not including the network of typewriter consoles surrounding each computer. Attempts at computer networks have been made in the past; however, the usual motivation has been either load sharing or interpersonal message handling. Three other more important reasons for computer networks exist, at least with respect to scientific computer applications. Definitions of these reasons for a computer network follow."
"One way to make the implementation of a network between a set of time-shared computers more straightforward and unified is to build a message switching network with digital interfaces at each node. This would imply that a small computer, an interface message processor (IMP), would be located with each main computer to handle a communications interface. It would perform the functions of dial up, error checking, retransmission, routing and verification. Thus the set of IMP's, plus the telephone lines and data sets would constitute a message switching network
"The major advantage of this plan is that a unified, straightforward design of the network can be made and implemented without undue consideration of the main computer's buffer space, interpret speed and other machine requirements. The interface to each computer would be a much simpler digital connection with an additional flexibility provided by programming the IMP. The network section of the IMP's program would be completely standard and provide guaranteed buffer space and uniform characteristics, thus the entire planning job is substantially simplified. The data sets and transmission lines utilized between the IMP's would most likely be standardized upon, but as changes occurred in the communication tariffs or data rates, it would be more straightforward just to modify the program for the IMP's rather than twenty different computers. As soon as the need became apparent, additional small computers could be located at strategic connection points within the network to concentrate messages over cross-country lines. Finally, the modifications required to currently operating systems would be substantially less utilizing these small computers since there would be no requirement to find buffer spaces, hold messages for retransmission, verify reception of messages and dial up telephone lines."
At this meeting, Larry Roberts meets Scantlebury, Davies, and Baran. [Roberts, Net Chronology] [ISOC] [Abbate p 38]
David Walden joins BBN and Frank Heart's Division. [Walden 1990 p 5]
| Bob Taylor (internet visionary) talks at UT Austin
April: Kleinrock receives an ARPA research contract. The ARPA funding went to support PhD students such as Cerf, Crocker and Postel. This was 100% of Kleinrock's funding at UCLA [Babbage 6, 14, 15] [Roberts, Net Chronology]
June 3 ARPAnet Plan "Resource Sharing Computer Networks" submitted for approval. [Salus p 26]
June 21 ARPANet plan is approved by Robert Taylor with a budget of $2.2 million. [Salus p 26] [Abbate p 56]
ARPA RFQ for IMPs is released Aug., drafted by Larry Roberts and Barry Wessler. [ISOC] [Roberts, Net Chronology] [History of Telenet p 29]
September: Twelve proposals were received. [Hauben] [Roberts, Net Chronology] [Walden 1990 p 5 (helped prepare BBN's bid)]
"It was an extremely difficult problem to produce a proposal like that in a very short time, because the government has the very natural tendency to spend an enormous amount of time getting an RFP out. Then they breathe a big sigh of relief, and they want the people to write the proposal in zero time." Frank Heart 1990.
The other finalist bid for the ARPA contract was from Raytheon. [Nerds2.0 p. 79] [Vanity Fair]
Jacobie Systems in Santa Monica was a bidder [Cerf, Oral History 1990] Steve Crocker and Vint Cerf worked for Jacobi Systems. [Cerf, Oral History 1990] [Vint Cerf and Steve Crocker, Nov. 2011 Smithsonian American Art Museum lecture]
IBM and AT&T did not bid on the contract [Segaller 62] (note that AT&T may have been prohibited to have built the IMPs pursuant to the MFJ) IBM apparently considered the project "impossible" and too expensive. Two decades later IBM had learned from its mistake and was making great concessions in order to be a part of NSFNET. [Segaller 72] [Roberts, Computer Science Museum 1988] ("IBM and CDC 'no bit' it because, they told me: "If we used Model 50s at these sites, you're going to go broke, and we think you're crazy." ")
JCR Licklider and Robert Taylor, The Computer as a Communications Device, Science and Technology 1968 [Markoff Dec. 20, 1999]
December: Elmer Shapiro, A Study of Computer Network Design Parameters, SRI ("Ideally, the communication subsystem should represent a 'transparent pipe' between hosts, in that a message from host A for host B would arrive at B as an identical copy of the source message.") [Salus p 25] [Def Tech Info Center]
L. ROBERTS "The ARPA network" Invitational Workshop on Networks of Computers Proceedings National Security Agency 1968
USG initiates antitrust lawsuit against IBM.