MORE ONLINE

Last month we looked at the short history of the World Wide Web, which is the computer program that made the Internet accessible to ordinary people instead of just computer technicians and hobbyists. This month, we shall look back even further, almost 40 years, for a short history of the Internet itself. You thought it was younger than that? Well, it doesn’t show its age in a lot of ways, but some say it might be slowing down a bit.

The Internet is not a giant central computer, as some might imagine. There is not even a central computer which acts as a hub for Internet traffic, although Microsoft and Netscape would like to think so. So what is it? The Internet is the connection of many computers, all over the world, primarily through telephone wires, but also by cable, satellite and radio relay. The Internet uses a special kind of connection, different from the direct connection used for telephone and facsimile (fax) transmissions. Computers on the Internet are not connected by a direct line like the ones used for your telephone or fax. The direct connection used for telephone calls is switched on (connected) at each central exchange, so it’s called "circuit switching." The method used for Internet data transmission is called "packet switching."

The idea of sending large amounts of data from one computer to another over telephone wires, without having a direct connection between them, is an idea that still borders on science fiction. Some very brilliant electrical engineers thought about how to do this, at about the same time, in different parts of the world. The idea of sending small blocks of data over the telephone network was conceived almost 40 years ago by Leonard Kleinrock, and documented in his Master’s thesis at MIT in 1959. In July 1961, his article "Information Flow in Large Communication Nets" appeared in the RLE Quarterly Progress Report. In December 1962 he presented his Ph.D. thesis, "Message Delay in Communication Nets With Storage." This is the theory on which the Internet would be based.

When a client computer "inquires" data from a host computer on the Internet, the data is divided into conveniently sized "packets" which are sent over telephone circuits one at a time, and routed electronically to their destination, where they are reassembled into the document (or picture) that you requested. The packets may follow different routes on their way to your computer, so there is sometimes a slight delay in putting them back together. This delay would not be acceptable for voice transmission, but it works fine for email and for searching the Internet for information, which is exactly what it was designed for.

There are millions of computers connected as data servers on the Internet, so the disconnection of any one (or thousands) would not make any difference. This is an important point, and it has to do with why the word "Internet" is always capitalized. The server computers are not always online, except in a standby mode. Think of all the computers on the Internet as all the Aunt Marthas in the world, each one waiting for the phone to ring so she can tell you all about the goings-on in her neighborhood, and a bunch of other things that she overheard on the party line.

At about the same time that Len Kleinrock was earning his Ph.D. at MIT, the term "packet switching" was coined by Donald Watts Davies, a British computer pioneer. In the mid-1960s he described network concepts very similar to (but independent of knowledge about) Paul Baran’s work at RAND Corporation in the USA. Baran was working on a similar idea of breaking down messages into small packets, but for the purpose of secure military communications, not computer networks. Somewhere the idea was mistaken that the Internet was developed in order to make military communications secure in the event of nuclear war. That was never part of the specification written by the architect of the Internet, Lawrence Roberts of ARPA. However, Baran and Davies are recognized as co-developers of the packet-switching technology that is used in virtually all modern computer networks, including the Internet.

In contrast, plain old telephone service (POTS) is based on circuit-switching technology, in which an individual circuit is switched on (connected) for direct transmission between two telephones or fax machines through the network of telephone lines and switching centers. Circuit-switching is required when data must be transmitted quickly and must arrive in the same order in which it’s sent—in other words, in real time. This is necessary for voice and fax transmissions. Packet switching is more efficient, but it allows some delays in transmission, which is acceptable for email messages and for surfing the Web, but not for voice.

In the mid-1950s it became apparent that technology was the weapon of choice for the Cold War, and there were indications that the Soviets were winning. According to Tom Wolfe’s book, "The Right Stuff," when (then Senator) Lyndon Johnson was told that the Russians had beaten the U.S. in launching the first man-made satellite (Sputnik, October 4, 1957), he replied, "You’re telling me that their Germans are better than our Germans?" The Cold War had become a race for technology.

One of the answers to the problem of lagging technology in the U.S. was the creation of the Advanced Research Projects Agency (ARPA). An agency of the U.S. Department of Defense (DoD), it was charged with the development of new technology for use by the military. One of the scientists hired by the agency in 1962 was Joseph Licklider, an eminent psychologist and visionary computer scientist, to head ARPA’s behavioral sciences division. He would lead ARPA into the computer research which would eventually become the Internet. It’s interesting that two of the ARPA scientists who had key roles developing the Internet were specialists in the science of psychoacoustics (Licklider and Bob Taylor). Are there some vibrations out there that the rest of us don’t hear?

It’s also interesting that ARPA changed its name to Defense Advanced Research Projects Agency (DARPA) in 1971, then back to ARPA in 1993, and back to DARPA in 1996. In spite of its identity problems, DARPA is still engaged in high-tech research for the government. One of the current projects it is funding is the development of a six-inch spy plane that is essentially "a flying [micro]chip," according to its designers at MIT’s Lincoln Laboratory. Yes, that’s six inches. Well, back in the 1960s the idea of a computer network was regarded as just as screwy as the idea of a six-inch spy plane. Wait a minute—don’t those guys have an aerodynamicist on the staff there at Lincoln Lab? Why doesn’t someone tell them that a six-inch airplane with a supersonic shape would have to weigh less than a gram or two and, in order to fly, it would have to go so fast that it wouldn’t be a useful platform for aerial surveillance? They expect it to fly at "20 to 30 mph," according to an article in this month’s Technology Review. It might be easier to tame a hummingbird for this job.

ARPA’s office of computer technology (called Information Processing Techniques) was headed by Bob Taylor from 1966 to 1969. He came up with the idea of building an experimental computer network based on Len Kleinrock’s theories, and convinced the agency to provide funding of one million dollars to try it. Taylor arranged for Larry Roberts at the Lincoln Laboratory at MIT to conduct some experiments to prove the feasibility of such a network. Roberts had already been doing advanced computer research at Lincoln Lab. There, under his direction, two computers "talked" to each other for the first time in October 1965, when the Lincoln Lab’s TX-2 computer in Massachusetts was connected by a long-distance telephone call to the Q32 computer at System Development Corporation (SDS) in Santa Monica, California. This test hookup was a dedicated line, not like the packet-switching technology that would come later.

Eventually Roberts was persuaded to head up the network project at ARPA, where he wrote the specification for the ARPA network, or ARPANET, as it would be called, and drafted the Request for Proposals which resulted in their contracting Bolt, Beranek and Newman (BB&N), an engineering consulting firm in Cambridge, Massachusetts. BB&N was contracted to design the hardware and software for a data transfer network that would permit universities and government research centers to easily share information that was stored on their computers, using the existing telephone networks and packet-switching technology.

BB&N developed the first Interface Message Processor (IMP), based on a Honeywell mini-computer with 12 Kilobytes of memory. IMPs are now referred to as "packet switches," or simply "switches." There was still a lot of work to be done before the first packet data could be exchanged over the ARPANET. In addition to the hardware, a computer program had to be developed which would inquire data from a remote "host" computer, break up a data file into packets, code each one with the "address" of the inquiring computer, send them over the telephone network, and reassemble them into the complete data that was requested. This type of computer program, or software, is called a data-transmission protocol.

Vinton Cerf and Bob Kahn are credited as co-developers of the Transmission Control Protocol/Internet Protocol (TCP/IP), which are the rules of the road, so to speak, for the Information Superhighway. Bob Kahn worked on the IMP project at BB&N, and Vin Cerf was a graduate student in computer science at the University of California at Los Angeles (UCLA), where he helped to install the first IMP for packet-switched data transfer.

Vin Cerf and other graduate students worked at the UCLA Computer Science Department’s Network Measurement Center, under the direction of Len Kleinrock. ARPA chose UCLA’s Network Measurement Center to set up and conduct the initial tests of the network, and that laboratory became the first node of the ARPANET when BB&N’s first IMP was connected to their Sigma 7 computer in September 1969, just 28 years ago. A month later the second node of the fledgling network was Installed at Stanford Research Institute in Palo Alto, California, when Doug Engelbart’s group connected the second IMP to their SDS 940 computer. The first time that the ARPANET was tested, it went something like this, according to a recent interview with Dr. Kleinrock, which appeared in the Knight-Ridder newspapers:

Kleinrock and his small group of graduate students at UCLA planned to log onto the computer at Stanford University, over 300 miles away, and try to send it some data. They agreed to start by typing "login," and checking to see if the letters appeared on the distant monitor.

"We set up a telephone connection between us and the guys at SRI," Kleinrock, now 62, said in the interview.

"We typed the L and we asked on the phone, 'Do you see the L?' "

"Yes, we see the L," came the response from Engelbart’s group at Stanford.

"We typed the O, and we asked, 'Do you see the O?' "

"Yes, we see the O."

"Then we typed the G, and the system crashed…"

From this humble beginning sprang the Internet. In November 1969 the third node of the ARPANET was installed at the University of California at Santa Barbara (UCSB), and in December the fourth node of the ARPANET was installed at the University of Utah. In March 1970 the BB&N computer in Massachusetts was hooked to the ARPANET, completing the first nationwide computer network based on packet-switching technology.

In 1990, the Department of Defense began phasing out the ARPANET, and administration of the network was transferred to the National Science Foundation, a non-governmental agency funded by various scientific and academic organizations. Currently, the Internet is not under the regulation of any government agency. Under the NSF, the Internet is politically independent and self-funded through commercial aspects, primarily user fees and advertising. It is currently administered by InterNIC, an agency established in January of 1993 as a collaborative project between AT&T, General Atomics Corporation and Network Solutions, Inc. The latter was the first agency established by the NSF to administer the Internet after ARPA administration was ended. Under the 1993 agreement, AT&T was to manage the Internet Directory and Database Services project; Network Solutions would continue to manage the Registration Services project, and General Atomics was to manage the Information Services project.

The agreement stipulated that a peer review during the second year of performance would determine the future level of funding. The review found that General Atomics was not providing the promised level of service to the community and recommended that funding be discontinued. NSF agreed with this recommendation and, in February 1995, terminated the cooperative agreement with General Atomics. Steps were taken to minimize the disruption to the research and education community and to continue the services which the panel identified as having significant value.

Although the U.S. Department of Defense has ceased to administer or fund the Internet, there is another U.S. government agency charged with coordinating other government agencies’ use of the Internet. This is a user group, and has no administrative function relative to the Internet itself.

The Federal Networking Council (FNC) is chartered by the National Science and Technology Council’s Committee on Computing, Information and Communications (CCIC), to act as a forum for networking collaborations among Federal agencies to meet their research, education, and operational mission goals and to bridge the gap between the advanced networking technologies being developed by research FNC agencies and the ultimate acquisition of mature version of these technologies from the commercial sector. On October 24, 1995, the FNC unanimously passed a resolution defining the term "Internet." This definition was developed in consultation with members of the Internet and intellectual property rights communities. The FNC includes representatives of the following U.S. government agencies: Department of Defense (DoD), Department of Energy (DOE), DARPA, NSF, National Aeronautics and Space Agency (NASA), and Health and Human Services (HHS). Here is the resolution:

Resolution: The Federal Networking Council (FNC) agrees that the following language reflects our definition of the term "Internet."

"Internet" refers to the global information system that:

(i) is logically linked together by a globally unique address space based on the Internet Protocol (IP) or its subsequent extensions/follow-ons;

(ii) is able to support communications using the Transmission Control Protocol/Internet Protocol (TCP/IP) suite or its subsequent extensions/follow-ons, and/or other IP-compatible protocols; and

(iii) provides, uses or makes accessible, either publicly or privately, high-level services layered on the communications and related infrastructure described herein.

The Internet has changed very much since it came into existence almost 30 years ago. It was conceived in the era of large mainframe computers, for the purpose of academic and military research, but has survived into the era of personal computers and local-area networks (LANs). It was designed before LANs or intranets existed, but has accommodated that technology, as well as more recent developments such as asynchronous transfer method (ATM) and frame-switched technology. It was envisioned as supporting remote login, file sharing and scientific collaboration, and has spawned even more, such as electronic mail and the World Wide Web.

For those interested in a more technical history of the Internet, written by several of its inventors, see A Brief History of the Internet, written by Barry M. Leiner, Vinton G. Cerf, David D. Clark, Robert E. Kahn, Leonard Kleinrock, Daniel C. Lynch, Jon Postel, Larry G. Roberts and Stephen Wolff.

And if you’re interested in reading more about these gentlemen, let me recommend a whole book about them: Where Wizards Stay Up Late, by Katie Hafner and Matthew Lyon. It’s a fascinating story—almost as exciting as "The Right Stuff." Let’s see, do you suppose we can get Tom Hanks to play Larry Roberts?

During the week, Thomas More writes rather technical stuff about business software, but on weekends he likes to loosen up a bit and write about WWWiz stuff.