The Public Switched Telephone Network (PSTN)
The Advanced Research Projects Agency (ARPA)
Data Packets and Computer Networks
Intergalactic Computer Network
Interface Message Processor
E-Mail
Unified Protocol TCP/IP
Domain Names
ARPANET becomes Internet
The World Wide Web

History of the Internet

The Public Switched Telephone Network (PSTN)

So far in mobile communication (and in communication in general) focus has been primarily about the transmission of one thing: speech.

The telephone was invented to carry speech over longer distances. Initially only between two participants, for example in different rooms. Soon switching centers were developed in which operators made connections manually. Later those operators where replaced by mechanical switching systems.

There were local networks (for local calls) and national networks (long distance calls). Special connections have been established between larger cities and those connections have continuously improved and expended. Until the late sixties everything remained electrical and mechanical. Since then, with improved electronics, multiple calls could to be transmitted over a single line by modulation using several carrier frequencies. This was Frequency Division Multiple Access (FDMA). Finally digital transmission was introduced and allowed speech to be converted into digital signals, which could then be transmitted in a “time slot”. This started Time Division Multiple Access (TDMA).

While the lines to the actual telephones remained analog, the telephone network became more and more complex and, above all, more and more digital. It was completely controlled by software and a well defined protocol was defined to manage the network. This network was called the Public Switched Telephone Network (PSTN).

At the end of the 1980s, the industry was ready to digitize the entire network. The Integrated Service Digital Network (ISDN) was created. Here it was possible for the first time to provide digital lines to the end device, however only with a low transmission rate (2 x 64 kbit/s). The digital mobile network GSM adapted seamlessly into this communication structure. However, the central element of the network was still to provide a line between two participants so that they could communicate. These were mainly people who were connected via speech.

This principle was greatly underlined by the slogan of the Nokia: „Connecting People“.

But „establishing a line“ is not the only way to communicate. The goal couldn’t just be to connect people. Parallel to the development of powerful computers, a different, new type of communication emerged: Packet Switching.

Packet Switched Network

The Advanced Research Projects Agency (ARPA)

In 1957, the Soviet Union launched its first satellite, Sputnik. This was an alarming wake-up call for the United States. It was possible to send rockets anywhere and cause great destruction. The US was vulnerable in its own country. A particularly frightening scenario was that the entire telecommunications infrastructure could be destroyed. After the war, people were already 100% dependent on reliable communications for the country’s defense. However, as described above, the network was based on line switching and centrally organized. If, for example, a large long-distance transmission network was disrupted or even destroyed, nation wide communication would collapse.

Would it be possible to create a less vulnerable communication system? To investigate this and other pressing defense problems, the United States founded ARPA (Advanced Research Projects Agency). This agency was intended to promote and coordinate research that would contribute to improving U.S. defense. One of the ARPA projects focused on secure and attack-resistant communication networks.

Data Packets and Computer Networks

One of the first engineers to think about a new type of communication was Paul Baran. He worked for a non-profit organization called RAND that focused on the vulnerability of communications systems. Baran proposed a complete new communications network. This consisted of network nodes and connections between the nodes. The following figure shows an example of a network of 12 such nodes (A – L).

Each node is connected to a number of other nodes. For example, if node A wants to communicate with node K, this can happen via multiple paths. Three different paths are shown in the figure. Communication requires a bundle of data to be sent and work its way from node to node. The nodes must be “intelligent” and know how to forward the bundles. This makes such a network quite complex. On the other hand, communication works even if several lines or nodes are lost. The bundles can practically find a way around the defects.

In 1960, Baran used computer simulations of such networks and was able to show that when there were more than three or more connections from the nodes, the networks became very robust. They were still working even even if 50% of the lines and nodes failed.

In the following years, Baran worked on the specification of how such a network could be set up and published the results in 1964. His work and suggestions were rather laughed at by the large communications companies such as AT&T. “Baran simply has no idea how communication works in reality.” Speech communication would never work this way.

In the early 1960s, a completely independent scientist in England was working on the same problem. He was not an electrical engineer like Baran, but a computer scientist. His name was Donald Davis. He immediately called his network a computer network. His main concern was the networking of computers.

Davis‘ focus was on the data that should be transferred. He suggested dividing the data into small “data packets”. Each of these packages was given an address to which it should be sent and a sender from where it was sent. Furthermore, the number of the respective package and of course the actual data. When arrived at the destination, the package would “unpack” and the data would be reassembled. Each data packet could independently find its way through a computer network, similar to what Baran had described.

Davis published his work in 1966. His term for data transmission using packets was to become established worldwide. The data transfer took place on computers that had to process a protocol for this. He recognized early on that processing the protocol and passing on data and information through the computers created a certain burden and would quickly become a limiting factor. He foresaw that there would be special computers that specialized in these tasks. These special computers, which he called “interface computers,” are now known as routers.

INTERGALACTIC COMPUTER NETWORK

One of the most colorful people in the early days of computer science was J.C.R. Licklider. In the early 1960s he worked for a well-known research and development company BNN (Bolt Beranek and Newman Inc). Although he actually came from acoustics research and was even President of the Acoustic Society of America at that time, he was interested in computers and their impact on the future of humanity from an early age. During his time at BNN, he wrote articles about a possible network of computers. He gave this series the strange name “Intergalactic Computer Network”. His vision was that computers would store and process information that would be accessible to everyone, the government, institutions, companies and even individuals.

In 1962, Licklider also worked for ARPA as director of the Information Processing Techniques Office. Here he had the opportunity to promote his idea for a computer network. He convinced two scientists, Ivan Sutherland and Bob Taylor, to work on such a network. He himself left the ARPA in 1964 to devote himself to new tasks. Sutherland and Taylor, in turn, began activities for a first computer network, which they called ARPANET. During their investigations, they soon became aware of the activities of Davis in England and Baran in the USA and incorporated their ideas about a packet-oriented computer network into the design of the ARPANET.

INTERFACE MESSAGE PROCESSOR

The aim of the ARPANET was initially to connect a handful of computer. Today this sounds like a simple, obvious task. But in the 1960s, computers were islands and in no way designed to exchange data among each other.

A computer at that time was working on a single task, a „job“. It loaded the corresponding program from either a magnetic storage, a magnetic tape or a stack of punch cards and executed it. The result was typically printed. A program was usually available in the form of a stack of punch cards. This was placed into the computer’s punch card reader and read. This is why it is referred to as „batch operation“. One job after another was executed in a row. Computers were not able to run multiple jobs at the same time (time sharing). However, this kind of time sharing was necessary if computers were to communicate with each other.

Computer in the sixties and seventies could only operate on a single task. This made it impossible to execute „network tasks“ as a side job.

The ARPANET team quickly came to the conclusion that established computers could not communicate directly with each other. Therefore, they proposed a different architecture. Each computer to be networked was connected to a second smaller computer. This was then responsible for exchanging data. It basically run a single task „communicate with other computer“. They called this computer IMP (Interface Message Processor). In mid-1968 there was a specification of what the IMP should look like and a search was underway for a suitable manufacturer. BNN, the company that Licklider worked for for a while, was awarded the contract to build and commission the IMPs.

BNN selected a so-called minicomputer for the IMPs, which was built by the Honeywell company. „Minicomputers“ was quite misleading since the devices were still as big as refrigerators. But they had a small word size (usually 16 bits), a very small RAM (24 k bytes) and a simple instruction set. Each IMP could communicate with up to six other IMPs via modems. Telephone lines were used for this and had a transmission rate of 64 kBit/s. At that time, 64 kBit/s transmission was a new, innovative digital transmission that was just introduced by AT&T Bell Labs. As described above, 64 kbit/s was sufficient for a voice channel. BNN also created the first software to control the transmission. It was simply called 1822 because it was described in BNN Report No. 1822.

Of the first four IMPs, three were installed along the California coast and one in Salt Lake City, Utah. The first transmission between two IMPs took place on October 29, 1969 from Stanford to Los Angeles. The letters “lo” from „Log In“ were received. By December 5th all four IMPs were connected. The first initial ARPANET was in operation.

December 1969 onwards, the ARPANET grew slowly and steadily. By 1975, 60 computers in the United States were already connected. New, constantly improved protocol software such as the Transfer Control Protocol (TCP) emerged. The ARPANET project had served its purpose but the network continued to exist. BBN released the software for the IMP and the HW and SW of IMPs could also be further developed. Unlike developments in the telecommunications industry, the ARPANET was characterized by research by the affiliated universities and institutes and the exchange of knowledge.

E-Mail

In 1971, another element for exchanging data, e-mail, was created in the ARPANET. As early as the 1960s, it was possible for users to send short messages to each other within a single computer. To do this, they wrote a message to the addressee’s file using their username. A programmer named Ray Tomlinson expanded this system by using the ARPANET to send messages to other external computers as well. Here he had to expand the address to include the user name plus host name, i.e. the corresponding computer in the network. He looked for a „separating character“ and chose a little-used ASCII character: @. Over time, this symbol developed not only as a symbol for e-mail, but as a symbol for the Internet itself.

Unified Protocol TCP/IP

ARPANET was the leading, but not the only packet-oriented network to emerge in the 1970s. There were other smaller networks in Europe too. What was missing was a common protocol that connected everyone together. Over the years, the so-called Transfer Control Protocol TCP and Internet Protocol IP developed. TCP was first demonstrated in 1977 and then further refined. In the early 1980s, TCP/IP was finally declared a standard by the U.S. Department of Defense. On January 1, 1983, all ARPANET nodes were converted to TCP/IP. Since then, TCP/IP has been the basic protocol for the Internet to this day.

Domain Names

Packet transmission need IP addresses. IPv4 addresses consisted of 4 numbers from 0 to 255 separated by a dot. Example: 192.0.2.42. (Today the addresses with IPv6 are significantly longer). However, such addresses are difficult to remember. Therefore, a database was created in which IP addresses were easily assigned to common names. These were called domain names and the Domain Name System (DNS) database.

On January 1st, 1985, nordu.net was the first domain on the Internet.

ARPANET becomes the Internet

At the end of the 1980s, the ARPANET was disbanded and became an Internet backbone. The military part of the ARPANET separated. In 1990 the network was finally released for commercial purposes and now left the military or research and development sector.

The lines that were previously used for the network came mainly from the telephone companies. These improved significantly in the 1980s by allowing higher bandwidths. Ultimately, cables were also laid specifically for the Internet backbone. Fiber optic cables were increasingly being used instead of copper cables, which not only allowed higher transmission rates, but could also send signals with low attenuation over long distances. In 1988 the first overseas fiber optic cable was laid. By the mid-1990s, fiber optics had already replaced copper cables in long-distance transmission.

Until the 1980s, IMPs were minicomputers with special software. In 1985, a company was founded out of Standford University that specialized in building IMPs, which were henceforth called routers. That company was CISCO. It quickly established itself as the market leader in routers and had incredible growth in both sales and value. At the end of 2000, at the peak of the Internet boom, CISCO was the most valuable company in the world with a market value of 500 billion.

In the 1980s, a new type of computer, the personal computer appeared on the market and established itself as a consumer product. Initially it was dominated by hobbyists interested in computer technologies, but it soon became an object of everyday use. With the PCs, the first modems also developed, which made it possible to connect to a network.

So-called Internet service providers were established that offered services to users. These were the first e-mail services but also access to newspapers. Current stock market values were also popular. CompuServe and America Online were the first major ISPs on the market.

The World Wide Web

The most popular application for the young Internet was again developed from research. This time it was actually not coming from the USA but from a research facility for nuclear physics in the French-Swiss border area, CERN. In 1989 researchers wanted to create a better system for exchanging information. This was created by the Belgian scientist Robert Cailliot and the British physicist Tim Berners-Lee. It consisted of a hypertext language called HTML (Hypertext Markup Language) and a protocol HTTP (Hypertext Transfer Protocol). It was possible to design information pages using HTML and transmit them using HTTP. What was special were so-called links. It was possible to click on such links to get directly to an other HTML page.

The first website on the Internet was created on December 20, 1990. Required software and protocol were license-free and available to everyone. Soon the WorldWideWeb developed also outside CERN and related research institutes.

However, the World Wide Web only gained great popularity through the development and commercialization of so-called browsers, i.e. software that facilitated the use the World Wide Web in a simple way. These browsers ran on the PCs in people’s homes and the World Wide Web became practically synonymous for the Internet even though it was actually just an application of the Internet.