ARPANET

The new awareness of the potential of technology caused scientists and the military to begin examining their technology and infrastructure, especially voice communications. The fear was that if the Soviets could launch a rocket that puts a satellite in orbit, they could probably also launch a missile that could conceivably take down the telephone communication system in the United States. What was needed, they projected, was a system that could withstand an attack and remain operational.

One solution proposed was put forward in 1962 by an MIT scientist, J.C.R. Licklider for a “galactic network” of interconnected computers that could communicate to one another. Rather than replace the telephone system, this network would be used should the Soviets destroy the phone system. While this proposal generated interest, the technology to make the network work had yet to be fully developed.

MIT scientist Paul Baran and Donald Davies, a scientist at the National Physical Laboratory (NPL) in England, independently developed a means to transmit “packets” across a switched network of computers. Baran called his “distributed adaptive message block switching,” but it was the name Davies used that caught on—”packet switching.” Packet switching was the key to creating a computer network that could not be taken down by a single site being destroyed.

In 1966, Licklider and a team of developers designed what became known as the ARPANET project. ARPAnet was completed and ready for testing in October 1969. So three months after Neil Armstrong walked on the moon, Charley Kline, a student at the University of California at Los Angeles (UCLA), was to send the message “LOGIN” to a computer at Stanford University. However, the network failed after only the “L” and the “O” were entered and received at Stanford. An hour later, Kline was able to enter the entire word, which Stanford received. While this test only confirmed the transmission of data between two computers connected point-to-point, it also proved the concept of the Interface Message Processor (IMP), the early precursor to a router. By the end of 1969, the University of Utah and the University of California at Santa Barbara (UCSB) were added to ARPAnet and packet switching was first used.

The ARPAnet continued to add nodes, as shown in Table 1-2. Growth started slowly, but it gained momentum as the network and its new technologies proved to be beneficial. The University of Hawaii was added, and the network went international in 1971, adding nodes in England and Norway. The ARPAnet had grown into an internetwork, which is a network of networks that are interconnected for communication. The term internetwork was eventually shortened to internet. Figure 1-4 shows an early sketch of the architecture of the ARPAnet.

By the end of the 1970s, the network began having problems with the number of IMPs that shared a single network framework. In the early 1970s, two computer scientists, Vinton (Vint) Cerf and Bob Kahn, developed the Transmission Control Protocol (TCP) that originally included the functions of what was later separated as the Internet Protocol (IP). These developments solved most of the communication problems by forming the distributed nodes of the internetwork into smaller groupings that better facilitated communication.

In the early 1980s, the Network Control Protocol (NCP) that had been the mainstay of the ARPAnet was replaced by the TCP/IP suite of protocols. There is no doubt that TCP/IP made a worldwide network possible. What was limited to primarily education and government during the 1980s began to be attractive to medicine, commerce, and industry.

The Legacy of ARPANET

Much of what we now associate with the internet was developed under ARPAnet. Here is a list of some of the important technologies and tools originally developed for or because of the ARPAnet:

• Electronic mail (email) (1971)—Ray Tomlinson developed a program that sends electronic messages over a distributed network. Limited to the characters available on Teletype devices, he chose the at-sign (@) to separate local addresses from global addresses.

• Telnet (Teletype network) (1972)—Developed originally for NCP, telnet provided a bidirectional communications link between two stations that allowed a local device to work at the command prompt of a remote device.

• FTP (File Transfer Protocol) (1973)—Another protocol originally developed for NCP by Abnay Bushan, it facilitates the transfer of files and bulk data to be sent from one node to another.

• IS-IS (Intermediate System to Intermediate System) Protocol (1980s)—Radia Perlman developed a method to route IP messages. Later, she developed the Spanning Tree Protocol (STP) to eliminate switching issues on Ethernet networks.

• Domain Naming System (DNS) (1983)—Paul Mockapetris and John Klensin develop DNS, which allows the network to expand beyond education and government.

• Open Systems Interconnection (OSI) Reference Model (1984)—What is now commonly referred to as the OSI model was issued jointly by the International Standards Organization (ISO) and the International Telecommunication Union (ITU) as a specification for a universal protocol suite. The OSI protocol suite was not successful, but the reference model has remained a standard for internetwork communication.

• NSFnet (1986)—The National Science Foundation (NSF) left the ARPAnet to form its own competing network (NSFnet). As this network grew from connecting five supercomputers to virtually every major university in the United States, many of the developments that are still in use today allowed the NSFnet to outperform the older and slower ARPAnet. NSFnet was one of the primary components of the internet as we know it.

By 1983, the Pentagon felt that the general lack of security and the growing number of nodes with network access (and some believe the movie War Games) presented too much of a risk and split off to form the MILNET with very strict access rules. ARPAnet continued to support its academic and nonmilitary government members.

Other complementary or competing networks also began to appear:

• Computer Science Network (CSNET)—A network developed to provide networking capabilities to computer science programs at academic and research institutions that were unable to connect to the ARPAnet because of funding or governance limitations.

• Energy Sciences Network (ESnet)—A network serving the U.S. Department of Energy (DOE) scientists and collaborators worldwide.

• NASA Science Internet (NSI)—The NSN and SPAN networks were merged to create an integrated communication platform for the earth, space, and life sciences. At its peak, NSI connected more than 20,000 scientists around the world.

• NASA Science Network (NSN)—Space scientists were connected to data stored around the world.

• Space Physics Analysis Network (SPAN)—SPAN provided communication for investigators and engineers and access to science databases.

There is little doubt that the ARPAnet and all of the scientists, programmers, and visionaries who contributed to its successes and failures, from which much was learned, effectively built the foundation for the internet of today. The NSFnet and the smaller networks that were eventually absorbed into it replaced the ARPAnet in the late 1980s and was itself shut down in 1995. The control and management of the national and international internetwork became privatized, which opened the network of networks to commercial use.

The Maturing Network

One new development often leads to one or more developments. This has certainly been the case over the history of the internet. However, one series of developments has had perhaps the most impact on the internet since the IMP: the TCP/IP protocol suite, hypertext links, and the World Wide Web. These developments changed and improved how we use the internet, increased its popularity, and facilitated its expansion.

TCP/IP Model

The Transmission Control Protocol/Internet Protocol (TCP/IP) suite is the standard set of protocols for connecting and interconnecting computers and networks into the internet. Essentially, the TCP/IP protocols are the working parts that make the internet work. To understand what TCP/IP is and how it works requires that we look at the TCP/IP Reference Model and the protocols on each of its layers. As shown in Figure 1-5, the TCP/IP model has four layers (top to bottom):

• Application Layer—This layer of the TCP/IP model is where a host computer or a server interacts with a network to create or interpret the initial or final component of a communication. The name Application refers to protocols and services, such as email, FTP, and Telnet clients, which interact directly with other protocols. The functions that operate at the Application Layer include the identification of senders and receivers, access to remote hosts, and the request and retrieval of content from other hosts and servers.

• Transport Layer—Much like the functions of a shipping and receiving department, the Transport Layer of the TCP/IP model packages or unpackages outgoing or incoming messages, determines how they are to be transported, and monitors their transit. At this layer, functions such as flow control, error control, and segmenting or desegmenting are performed.

• Internet Layer—This layer of the TCP/IP model, which is also known as the Network Layer, performs the essential tasks involved with physically moving messages from the sending host or server to the receiving host or server. In performing these tasks, the Internet Layer applies addressing, forwarding, and routing to ensure messages arrive at their destinations, regardless of the path they may actually take.

• Link Layer—Also referred to as the Network Interface or Network Access Layer, the Link Layer provides the appropriate packaging of a message before it is placed on the network’s physical medium. Incoming messages are interpreted and formatted before being passed on to the Internet Layer. On a local network, this layer also provides physical-level addressing and control.

Hypertext

Hypertext provides internet users with the ability to move from one document to another on a network by clicking a link. It’s probably hard now to imagine the internet without hypertext links. Hypertext has become an assumed and required element of the internet fabric. As illustrated in Figure 1-6, a link embedded in one document can allow the reader to use the link to jump to related content on another document. However, unlike the majority of internet technologies, hypertext was developed over a period of about 60 years. The developers who contributed to what we know as hypertext today are covered in the following sections.