37
The short history of the Internet sketched in this section raises several questions
regarding the Internet development, questions that are worthwhile to explore more
profoundly in order to understand the characteristics of the Internet that shall play a
role in the competition policy analysis presented in this thesis:
• How did Internet connectivity, the fact that many thousands of independent
networks can be reached via one home network, evolve? This question will be
taken up in section 3.2 which looks at the development of computer communication and the development of connectivity in more detail.
• How did the present hierarchy between ISPs evolve? This aspect of Internet
history is looked into in section 3.3.
• Where did the common standards for Internet communication come from, and
who is responsible for the further advancement of Internet standards? This question is considered in section 3.4, which takes a closer look at the development of
the TCP/IP protocol suite and Internet standardization.
• Lastly, how did the terms for Internet interconnection evolve? This question is
taken up in section 3.5 which deals with Internet interconnection agreements.
3.2 History of computer communication and Internet transport services
The foundations for computer networking and ultimately for the Internet were laid
by the advent of computers and subsequent research into communication between
computers for purposes of remote login and time-sharing. Section 3.2.1 gives a more
detailed background on the motives for computer communication. Section 3.2.2 then
traces the development of computer networking to the Internet by discussing how
connectivity developed from limited interconnectivity in early computer networks to
the universal connectivity of the Internet.
3.2.1 Computer networking
The first computers were invented in the 1930s and 1940s. Despite their modest
processing power, these large machines filled whole rooms and were very expensive. It was therefore common, that they were shared by numerous users, when they
first became available for academic applications. In the 1950s, this resource-sharing
meant that a user would hand in a program for calculation to an attendant and receive the results after hours or days, depending on how many people were sharing
the computer, and on how complex the calculations were that were run. If a program
had a “bug” in it, an error or flaw in either the source code or the software design,
then it could take many iterations of this very time-consuming process to get good
results.
The limitations of this type of batch-processing motivated research into enabling
closer human-computer interaction, an effort that resulted in so-called time-sharing
38
systems, which were used at various computer science departments in the 1960s (see
Naughton, 2000: 73ff.). With time-sharing, several terminals attach to one mainframe computer. The machine cycles between the users that are online and allots its
entire processing capacity to one user for a short time interval. Because the machine
cycles fast, the user has the impression of having exclusive access to the resource.
Time-sharing made first real human-computer interaction possible, because a user
could debug and improve programs through a trial-and-error process. Time-sharing
systems also furthered the interaction in the community of computer users, since
users could help each other with programming by sharing files and memos on the
common system.
In 1962 the Advanced Research Projects Agency of the U.S. Department of Defense founded an Information Processing Techniques Office (IPTO) with the objective of funding basic seed research in computer sciences. The IPTO placed great
emphasis on time-sharing projects in the academic community (see Naughton, 2000:
81). Eventually, the incompatibilities between the different time-sharing systems
supported by IPTO at various university sites led IPTO to support research into
computer-networking that would allow the researchers at ARPA-supported sites to
access computer resources located at other sites from a single terminal (as opposed
to the existing terminals, which were configured to provide login to only one specific mainframe computer).
ARPA started a project on computer networking in 1966 (Naughton, 2000: 82ff.).
The result of this project was the ARPANET. This network consisted of specialized
mini-computers, which were interconnected by dial-up telephone lines that acted as
intermediaries between their attached computers (hosts) at the ARPA-funded research sites. The responsibility for the compatibility between the hosts and for
handling the communication between the intermediary computers was borne by
these intermediary-computers, which were under central ARPA control and could
thus communicate among one another via a standardized protocol. The communication between host and intermediary was the responsibility of each individual site. By
the end of the year 1972 the ARPANET had thirty-seven network nodes (meaning
intermediary-computers) (Naughton, 2000: 152).
While the ARPANET was the largest networking project supported by ARPA, the
agency also invested resources into the development of packet-switched terrestrial
radio networks and satellite networks. Because of its experience with multiple
packet-switching networks ARPA became interested in the possibility of enabling
communication between packet-switching networks of different underlying technologies (Leiner et al., 2003: 4). The ensuing research into a communications protocol
that could interconnect multiple networks with differences in technologies and
communications protocols resulted in the specification of the TCP/IP protocol.
The usefulness of computer communication and the possibility of sharing computer resources via remote login quickly carried the development of computer networks beyond the small community of ARPANET researchers. By the mid 1970s
computer networks were being installed by several federal agencies and research
communities. The U.S. Department of Energy, for instance, funded the MFENet for
39
research in Magnetic Fusion Energy and the HEPNet for its High Energy Physicists
(Leiner et al., 2003: 8). The National Aeronautics and Space Administration
(NASA) established SPAN, and the NSF approved a five year grant to finance the
CSNET, a network geared towards computer scientists of smaller universities without access to the ARPANET (Jennings et al., 1986: 946). A further network established in the early 1980s for interuniversity communication on the North American
East Coast was the BITNET.22 The USENET was developed as a simple store-andforward network for computers using the UNIX operating system.23 In 1988 the NSF
built the NSFNET, which connected its five supercomputer sites and offered longdistance data transportation between the attached local and regional networks.
There were also advances in computer networking outside of the United States.24
International networks based on packet-switching included the NPL network in the
U.K., where pioneering research on packet-switching had taken place. In France, a
research network named Cyclades/Cigale was started to test the U.S. and British
experiences using packet-switching for computer communication. A common European initiative was the European Informatics Network. ARPA collaborated with
researchers abroad, as for instance with the University College of London or the
Norwegian Defense Research Establishment, to establish direct links between the
ARPANET and international sites via satellite and ground connections (Hauben,
2004: 3ff.). Through these international collaborations, the ongoing work of refining
the TCP/IP protocols was always an international process. Many of the features of
later protocol versions, for instance, resulted out of discourses taking place at international informatics conferences.25
3.2.2 Developments in network interconnection
The early computer networks were at first designed and operated for closed user
groups. The networks enabled file-transfer and remote job entry in addition to communication via E-mail for users belonging to, for instance, a university consortium
or a government ministry. The use of these networks was tied to a common workenvironment and common projects. Some networks, especially those in the academic
22 Jennings et al. give a detailed overview of the other computer networks existing by the mid-
1980s (Jennings et al., 1986).
23 Naughton calls the USENET the “poor man’s ARPANET” (Naughton, 2000: 169). It was
based on the UNIX operating system, which was considered the most powerful operating system for mini-computers at the time. UNIX was distributed with open-source code and at low
cost and was therefore very attractive to academic users. On USENET each member-site covered its own costs for dial-up telephone line communication and agreed to forward packets of
other users within its own system.
24 See Hauben (2004) for information on the international origins of the Internet.
25 The biennial International Computer Communications Conference, which was started in 1972
in Washington, D.C., is such an institution where collaborative research relevant to the Internet development is initiated and discussed.
Chapter Preview
References
Zusammenfassung
Die Konvergenz der Netztechnologien, die dem Internet, der Telekommunikation und dem Kabelfernsehen zu Grunde liegen, wird die Regulierung dieser Märkte grundlegend verändern. In den sogenannten Next Generation Networks werden auch Sprache und Fernsehinhalte über die IP-Technologie des Internets transportiert. Mit den Methoden der angewandten Mikroökonomie untersucht die vorliegende Arbeit, ob eine ex-ante sektorspezifische Regulierung auf den Märkten für Internetdienste wettbewerbsökonomisch begründet ist. Im Mittelpunkt der Analyse stehen die Größen- und Verbundvorteile, die beim Aufbau von Netzinfrastrukturen entstehen, sowie die Netzexternalitäten, die im Internet eine bedeutende Rolle spielen. Die Autorin kommt zu dem Ergebnis, dass in den Kernmärkten der Internet Service Provider keine monopolistischen Engpassbereiche vorliegen, welche eine sektor-spezifische Regulierung notwendig machen würden. Der funktionsfähige Wettbewerb zwischen den ISP setzt jedoch regulierten, diskriminierungsfreien Zugang zu den verbleibenden monopolistischen Engpassbereichen im vorgelagerten Markt für lokale Netzinfrastruktur voraus. Die Untersuchung zeigt den notwendigen Regulierungsumfang in der Internet-Peripherie auf und vergleicht diesen mit der aktuellen Regulierungspraxis auf den Telekommunikationsmärkten in den Vereinigten Staaten und in Europa. Sie richtet sich sowohl an die Praxis (Netzbetreiber, Regulierer und Kartellämter) als auch an die Wissenschaft.