| The History of the Internet | |
| This was published in 1994 and is about the early days of what we now call the
Internet. It describes, among other things, the ARPANET, which,
although it was was created by the Defense Department, was conceived
of as a way to link researchers at various institutions.
There are links at the bottom to other histories of the Net, which are more detailed. You
are welcome to read these, but they are not required. From The
Internet for Everyone: A Guide for Users and Providers by Richard W. Wiggins |
|
| Internet Origins | |||||||||||||||||
| The Internet is often described as a "network of networks" The term is apt: there is no one authority that "owns" or administers the Internet; one could not draw a simple map of the Internet; there is not even a single authoritative list of the millions of computers or users with access to the Internet. Instead the Internet is a sort of confederation—a worldwide collection of national, regional, campus, and corporate networks. In order to understand today’s Internet, it will be helpful to consider its origins. | |||||||||||||||||
| The history of the Internet dates to the early days of computer networking in the
1960s. Today, when many workgroups have personal computers linked together in Local Area
Networks (LANs), it’s somewhat difficult to envision a world when mainframes
dominated, and getting those mainframes to talk to one another was a bold new concept. But
it’s important to realize that the beginnings and growth of todays Internet parallel
the evolution of computer networking itself. An agency of the U.S. Department of Defense, the Advanced Research Projects Agency, issues a request for proposals to link together four initial sites in July 1968. ARPA chose Bold Beranek and Newman’s proposal in December 1968. To this day, BBN remains an important provider of Internet-related technologies and services (in fact, a division of BBN was named in July 1993 to operate the regional network NEARnet). The initial sites were Stanford Research Institute, the University of California at Los Angeles, UC-Santa Barbara, and the University of Utah. In today’s terminology, we might call the network they were building a Wide Area Network, or WAN (albeit a rather small one). Just as there were no LANs and no PCs at the time, there were no interface cards to plug into the back of a computer in order to put it on a network. BBN built a set of Interface Message Processors, based on custom software and hardware on a minicomputer. The four sites were connected in the fall of 1969, marking the birth of the precursor to today’s Internet. The new network became known as the ARPANET. During this early experimental phase, early versions of some of the core protocols known to users today—including Telnet and the File Transfer Protocol—were born. A scheme called the Network Control Protocol managed the flow of data on the early ARPANET. |
|||||||||||||||||
| In these early days of the Internet, computers were much slower, communication lines
were much slower. . . and the number of computers on the network was small (only 200 hosts
by 1981). But even in these early days, the ARPANET design community realized that they
needed to build a network not of computers, but a network of networks. Various competing
networking technologies were evolving, and the ARPANET community wanted to be able to link
these disparate networks together. The basic protocols that support communications on
today’s Internet—TCP/IP--were developed in the mid-1970s by Vinton
Cerf of Standford University and Robert Kahn of BBN. TCP/IP
coexisted with NCP until 1983, when it replaced the Network Control
Program entirely..
The ARPANET grew as more installations--many of them defense department agencies, or research universities with ties to the defense department--became nodes on the network. While the ARPANET was growing into a national network, researchers at Xerox Corporation's Palo Alto Research Center were developing one of the technologies that would be used in local area networking--Ethernet. Over time, Ethernet became one of the important standards for how to implement a local area network. In the meantime, DARPA (renamed from ARPA) funded the integration of TCP/IP support into the version of the UNIX operating system that the University of California at Berkeley was developing. When companies began making powerful, independent workstations that ran UNIX, TCP/IP was already built into the operating system software, and vendors such as Sun included an Ethernet port on the back. TCP/IP over Ethernet became a common way for workstations to connect to one another. Throughout the 1980s, the corporate and university worlds were busy installing personal computers on the desktops of many of their professional employess. The same technology that made personal computers and workstations possible—processors and memory on inexpensive, mass-market chips—made it possible for vendors to offer relatively inexpensive add-on cards to allow garden-variety PCs to connect to Ethernets. Enterprising software vendors took the TCP/IP software from Berkeley UNIX and "ported" it to the PC, making it possible for PCs and UNIX machines to talk the same language on the same network. |
|||||||||||||||||
| By the mid-1980s, the Internet TCP/IP protocols were in use for numerous
campus-to-campus (wide area) links, and the same protocols were being used in many local
area and campuswide networks. The stage was set for explosive growth. FROM ARPANET TO INTERNET The term "Internet" appears in a planning document as early as 1974. But the network itself was called the ARPANET until the early 1980s. At that time, the U.S.Department of Defense decided to separate the military portion of the network into the "Milnet." The term "ARPANET" was still used for the portion of the network that was open to universities and corporate research divisions. At this point the ARPANET/Internet was still quite small, as the following chart shows:
|
|||||||||||||||||
| A milestone in the history of the Internet also took place in the mid-1980s. In
supporting five supercomputer data centers, the National Science Foundation (NSF)
concluded it needed a high-performance network linking those computers to one another, and
providing access to the machines for researchers across the country. In 1986 the NSF
solicited proposals to create a network linking the supercomputer centers. In 1987, NSF
proposed a newer NSFnet backbone, with much faster links between sites (1.5
megabits/second instead of 56 kilobits/second). The new backbone would link seven new
regional networks with the supercomputer sites. NSF awarded a contract to build and manage
the NSFnet to a partnership of IBM, MCI, and the Merit Computer Network. The creation of the NSFnet backbone and regional networks provided a milieu that encouraged Internet growth. A new data highway was in place, and many universities, keenly aware that leading research institutions had Internet access, wanted access for their own institutions. By the mid-1980s workstations either had built-in support for TCP/IP, or it culd be added inexpensively. At the same time that the marketplace was delivering millions of computers begging to be connected into campus networks, along with hardware and software to connect them, the NSFnet backbone and regional networks came into existence. Numerous universities decided that they needed to be on the Internet in order to compete and cooperate with the leading academic institutions. Of course, the academy was not alone in its interest in the Internet. Business and government were also installing personal computers and local area networks throughout the 1980s. TCP/IP was less pervasive in those settings; proprietary networks such as Novell, Banyan, or Appletalk, or mainframe architectures such as IBM’s SNA, were more common. Nonetheless, business and government represented pockets of growth on the Internet. Overall, the mid-1980s began a period of explosive Internet growth. This is not simply a U.S. story, of course. Academic institutions in Canada, Europe, and the rest of the world were not immune to these trends. . . .
There is every reason to believe this growth will continue. . . . The combination of the NSFnet backbone with the regional, or mid-level, networks is an important part of the growth picture. These networks provide services to the institutions and individuals of their respective areas of the United States. The idea of regional networks fits in nicely with the "network of networks" concept. The networks could be thought of as "mid-level" because in each case they themselves are networks of campus networks. The NSFnet backbone served as the high-speed link, connecting the supercomputer centers, other major research universities, and the regional networks. . . . ................................. The Web wasn't originally designed for commerce. Read "How the Web Was Wove" on the Business Week Online site.
Take a short quiz on the reading Supplemental readings: |