INET Conferences




Other Conferences

[INET'98] [ Up ][Prev][Next]

The Internet and Global Development

Michel ELIE <>
Observatoire des Usages d'Internet


The first part of the paper shows how network statistics can be used in order to stress some possible factors of growth of the Internet throughout the world. It examines the evolution of the geographical repartition of the Internet during the last two years on the basis of Network Wizards host distribution statistics.

The relation between the global distribution of Internet servers per unit of population and the level of development of each country as measured by gross domestic product (GDP) per capita or by the human development indicator (HDI) is shown, as well as the correlation between the level of development of countries, measured by the per capita GDP, and the number of Internet servers per inhabitant, which is somewhat similar to its correlation with the number of main telecommunications lines per capita put in evidence by the ITU in the classic Gipp curve.

The influence of other possible factors (such as the practice of English, type of government, insularity, and commercial dynamism) on the Internet situation in the different countries of the world is considered. In the future, the impact of the Internet on global development will also depend highly on the evolution of the Internet technology and of the services provided.

In the second part, another viewpoint, from the outside, is proposed to better understand the Internet relationship with global development based on the observation and study of the uses and practices of the Internet.

This implies collecting typical Internet uses which might differ from place to place, depending on culture and economics.

In conclusion, a worldwide network of observatories of the uses and practices of the Internet is strongly needed. The network would establish a set of basic public Internet statistics, observe its uses and practices, detect emerging ones, and facilitate the dissemination of uses which are profitable to the social, economic, and cultural development of the people. In France the "Observatoire des Usages d'Internet" is being started with this purpose.

This effort should be part of an international effort for more equity in the availability of the Internet tool worldwide, for the benefit of all, in order to keep developed and developing countries in the same planetary village.


1. Introduction

The impact of Internet growth on global development is probably still very difficult to understand. How and why the Internet is progressing in the world are also difficult questions to answer: to a great extent, the Internet development process is an enigma largely due to the unavailability of basic statistical data.

One would like to follow the real number of users, their characteristics, their geographic repartition, the use they are making of each service of the Internet and for which purpose and to know whether it changes the way of life of people, their income, their satisfaction in life, the organization of enterprise, the rules of society, etc.

There is a strong need for a better knowledge of the Internet development process in order to anticipate its technical infrastructure extension, to anticipate its impact on people and society, to take better advantage of this powerful and revolutionary tool for the benefits of our society, and to study corrections to the negative effects of a loosely controlled development process.

One of the objectives of this paper is to suggest different possible approaches to acquiring the knowledge of the Internet development process (1).

One can try to understand and anticipate the growth of the Internet from several viewpoints which proceed roughly from two different approaches, respectively designated as endogenous when the Internet is observed from the inside (number of hosts, data flows, data contents, etc.) and exogenous when it is observed from the outside (services provided, uses and practices, etc.).

From the inside one can look at the way the Internet infrastructure is developing in order to satisfy the user's needs and how it benefits from new technologies and the number and repartition of hosts as well as analyze the data flows, the contents available on the Net. From the outside one will try to understand why people need to use the Internet, what are the factors which generate the demand, which uses, generic or specific to a population of users, are developing, and which are the factors of resistance to Internet progression.

2. Endogenous approach: the Internet statistics

Working from existing public Internet statistics, one can consolidate some intuitive feelings as well as develop a new hypothesis about the main factors, positive or negative, which drive the development of the Internet. We are going to show the kind of information one can derive from the oldest Internet statistics publicly available (the Network Wizard "Internet host count per domain" survey) about the way the Internet is progressing worldwide. (2)

However, this only gives clues or tracks to be followed, not scientific certainty, for many reasons:

  • Incomplete statistics: In the present state of public Internet statistics, only sites belonging to geographic domains can be localized; the growing importance of nongeographical domain addresses makes it more difficult to associate a site address with its country of origin. In January 1998, 47% of the hosts counted in the Network Wizards survey are in one of the ".com", ".net", or ".org" domains, compared to 36% one year before. Nongeographical domains represent 60% of the Internet number of hosts growth between January 1997 and January 1998. The new Internet naming scheme will, of course, accelerate this tendency to mask the country of origin of the host;
  • The development of intranets which hide a big number of hosts from the measurement tools; and
  • The complexity of the reasons behind the Internet growth which suggests that we take several approaches before coming to some conclusions.

The exploitation of the Network Wizards figures concerning hosts belonging to geographical domains named "G-hosts" still brings some useful information and material to consider. G-hosts of the countries whose population in 1995 was exceeding 1 million inhabitants are studied, excepting the US which could not be included in the study as most US hosts do not use ".us" domains. However, the study can easily be updated as soon as the full geographical repartition of non-G-hosts is available. A change in the method used by Network Wizards for counting hosts leads to 1998 figures being slightly lower than what they would have been with the preceding method. However, this is probably of second order for our present purpose.

2.1. Worldwide distribution of G-hosts per class

So that we can better visualize the worldwide repartition of G-hosts among the world population, countries are classified between the following classes of G-host density defined on the basis of the G-hosts density in each country:

  • Class A: fewer than 0.1 G-hosts per 1000 inhabitants
  • Class B: from 0.1 to fewer than 1 G-hosts per 1000 inhabitants
  • Class C: from 1 to fewer than 10 G-hosts per 1000 inhabitants
  • Class D: more than 10 G-hosts per 1000 inhabitants

Figure 1 shows that in 1998, more than 68% of the world population lives in a country where there is less than one Internet G-host for 10,000 inhabitants.

Figure 2 shows that in 1998, 87% of the G-hosts are located in countries with more than 10 G-hosts per 1,000 inhabitants. This is probably much less than the reality because a majority of nongeographical hosts are located in class D countries. Also, this proportion has increased from 80% to 87% between 1996 and 1998, which means that the gap is extending.

These figures show that the Internet is strongly concentrated among a limited number of countries and a small fraction of the world population, located in specific areas of the planet as is shown by figure 3 which shows the repartition of the population of European (except ex-USSR countries), Asian, and African countries between G-hosts classes to which it belongs, as well as the corresponding figures in millions of inhabitants.

2.2. Internet and telephone extension as a function of countries' per capita GDP

As the Internet presently relies on the telecommunication infrastructure, it seems interesting to see whether one can find the same type of relationship between the Internet equipment of countries expressed in number of G-hosts per thousand inhabitants related to their per capita Gross Domestic Product (GDP): the number of main lines per thousand inhabitants of countries related to their per capita GDP known as the Gipp curve.

Figure 4 shows that there is a strong influence, quasi linear on logarithmic scales, between the Internet extension and the per capita GDP similar to the well-known one shown by the Gipp curve for telephone equipment (figure 5). (3)

However, there are some differences. For example, in the Gipp representation of figure 5, ex-USSR countries can be characterized as more equipped for the same per capita GDP than the other countries. With the Internet, things differ, although there is still a correlation between Internet position in a country and per capita GDP; as shown in figure 4, European ex-USSR countries are still better equipped than average, but Asian ex-USSR countries tend to be less equipped, in relation to their per capita GDP.

2.3. Internet extension: analysis of deviations

Figure 6 represents the relationship between the per capita GDP and the number of G-hosts in European countries, except ex-USSR countries. It shows clearly three groups of countries:

Highly developed countries of northern Europe where the Internet is more developed than average: Finland, Sweden, Norway, Denmark, Netherlands, Denmark, Great Britain. These countries share openness to the outside world, usual practice of English.

Highly developed countries of central and south Europe where the Internet is less developed than average: France, Germany, Italy, Spain, Greece. These countries at least share the fact that they do not have such a usual practice of English.

Slightly less developed Eastern European countries where the Internet is more developed than average, relative to what could be expected from their per capita GDP: Slovenia, Czech Republic, Hungary, Slovakia, Poland.

The Internet extension in various countries as measured by the number of G-hosts per thousand inhabitants can also be related to their level of development by using the Human Development Indicator (HDI) (4) published by the United Nations Development Program (UNDP), as shown in figure 7.

Among the 32 countries with fewer than 1 G-host per million inhabitants (which are not shown on figure 7), 21 are from Africa and none from Europe; nevertheless, a number of them are medium developed countries: Iraq, Vietnam, Gabon, El Salvador, Turkmenistan, Algeria, Syria, North Korea; Libya is even just in the group of countries of high development. In the following group of Internet less-equipped countries, according to their level of development, one finds the following countries: Tajikistan, Uzbekistan, Mongolia, Cuba, Tunisia, Saudi Arabia, Iran, Belarus. Some of the following qualifiers apply to several of these countries: no free circulation of information, little openness to the rest of the world, nondemocratic government, USA embargo, and no usual practice of English.

On the other edge, in the lower right corner of figure 7, one finds the Internet well-equipped countries according to their level of development: Estonia, South Africa, Lettonia, Namibia, Nicaragua, Zimbabwe, Zambia, Ivory Coast, Senegal; these countries share some of the following characteristics: free circulation of information, democratic government, openness to the outside world, and usual practice of English.

2.4. The case of a less developed country: Benin

The extension of the Internet in Benin was thoroughly analyzed in mid-1997 by Ken Lohento (5). It might be a typical situation of Internet extension in a developing country. Ken Lohento reports that in August 1997, there were around 1500 registered Internet mail addresses, hiding around 6000 users, that is, 0.12% of the population of Benin. About 70% of the sample studied are students or professors, and 17% work in the computer business. Only 14% are women. Electronic mail is the main service used, mostly for keeping in touch with distant friends or members of the family; the Web is used by only half of the users and forum is used by 10%.

An average Internet user in Benin has to overcome a number of technical obstacles: bad quality of the national telephone network outside the chief town of Cotonou and congestion on the narrow bandwidth available to leave the country (64Kbps to be extended to 128 thanks to the US "Leland Initiative"), electric power shortage, and effects of thunderstorms (lack of lightning conductors and generating sets). The development of the telecommunication network is critical. Old hardware and lack of competence for maintenance and network administration are also big problems. The user also meets a big economic barrier: only 12% of the users polled pay their costs by themselves. All these reasons restrain the development of genuine content and applications on the Internet in Benin and it is doubtful whether such countries can really catch up without international help.

2.5. Other "endogenous" data to be analyzed

There is a need for studies in the following directions:

Analysis of the Internet data flows: Information on the different flows of data, their volume, their coding, and their origin and destination is available but not collected.

Analysis of the contents of Web pages and documents accessible on the web, forum, mail: Such study has been carried out in order to determine the presence of the different languages on the Internet (one could also look for the effective use of the different languages). Two main studies give very different results: One by Alis Technology with support from the Internet Society uses sampling and gives a ratio of the use of English compared to French of 46.6, while the method based on using the AltaVista search engine developed by Daniel Pimenta and the Funredes organization (6), gives a ratio of only 17.6. Once more this stresses the need for a shared methodology in order to provide reliable public Internet statistics.

The analysis of data temporarily memorized in the different levels of cache could also give information on the data really called on by the user and on his search strategies for reaching it.

3. Exogenous approach: the Internet uses and practices

A completely different approach to understanding the influence of the Internet on economic, social, and cultural development is to study the uses of the Internet. This implies recording typical Internet uses which might differ from place to place depending on culture and economics.

Although marketing organizations are already looking for the individual user behavior wherever this user is a potential consumer, less attention is given to collective uses and practices which are nevertheless much more significant in terms of social impact.

The Internet users are the consumers of Internet services; the way they draw on these services associated with the Internet technologies they use is the origin of the data flow.

If an Internet practice is defined as a repetitive sequence of actions done by one user or a group of users acting in coordination, some of them implying the use of the Internet, a use of the Internet will result in the adoption of this practice by a category of persons. Collecting the Internet uses and practices will serve several purposes:

  • Allow for studies which could help to anticipate the development of the Internet;
  • Share the Internet uses and return of experience on uses thus reducing the learning curve for users starting new uses;
  • Cross-fertilize the use of the Internet in one domain by the uses in another domain; and
  • Get a better knowledge of the impact of the Internet on society and organizations.

Sharing uses and communicating about uses could be a new service on the Net. It is also possible to influence the development process by publishing on the Net some of the innovative uses collected. One Internet use emerging in one domain or one country could be useful to users in another one: this could prompt more cross-fertilization of uses among the Internet. It could facilitate the dissemination of uses which are profitable to the social, economic, and cultural development of people.

In France the Observatoire des Usages d'Internet, mostly funded by voluntary contributions, is being set up in order to create such a service.

5. Conclusion: toward a distributed observatory

Presently the knowledge of the Internet and its uses is scattered: We need to improve our understanding of them. There is little public exploitation of the available data as there is for other infrastructures such as telecommunication or transportation.

As we have shown, not only is the use of this very powerful tool still in the hands of a few countries (and probably a minority of inhabitants of these "info-rich" countries), but also there is very little public and reliable information about it.

There is also fear that it will be used as an instrument of domination by the well-off countries which push its development, on the others, as shown in recent news:

A coalition of developing nations, led by Egypt, India and Pakistan, is protesting a World Trade Organization proposal for an Internet "free trade zone," saying that such a development would reinforce the dominance of North American and European countries in the online world. The coalition is proposing that no decisions regarding the creation of a tax-free Internet trading zone be made until the problem of Western dominance of the Internet is resolved. Trade officials predict that negotiations on creating the free-trade zone will begin next year at the earliest.(7).

In order to avoid this danger, the Internet should be better understood, more transparent, and more multicultural.

A distributed Observatory of the Uses of Internet, such as the one started in France, could provide shared methodologies for a better intelligence of the Internet.

International bodies, as well as developed nations and developing ones which represent the majority of the future Internet users, could play an important role in reducing the gap, which is presently increasing, between those who can and those who cannot use the Internet. It is not up to the Internet to tell us where is the future of our society, but it is up to us to drive the Internet for it to be a useful tool for building the society we want to build.

6. References

(1) For an analysis of the Internet situation in July 1996 cf: M Elie; Internet et développement. Un accès à l'information plus équitable?; Futuribles; Nov. 1996; pp. 43-64

(2) Network Wizards Internet Domain Survey for Jan. 1996, Jan. 1997, Jan. 1998;

(3) Sources for statistics used in this paper: for National Gross Product per inhabitant (1994), and Human Development Indicator (1994), Worldwide Report on Human Development, United Nations Development Program, 1997; for Population (1994), and number of main telephone lines per 100 inhabitants (1994) Report on worldwide telecommunication development; International Telecommunication Union, 1995

(4) The Human Development Indicator (HDI) published by the UNDP is a composite indicator including several aspects of the development of a country: economics, through the per capita GDP; health and quality of living through the life length expectancy; and level of education combining level of alphabetization of adult and level of schooling of children.

(5) Lohento, Ken; Radioscopie de la connexion du Bénin à l'internet; ; Jan 98; and The Internet in Benin: an overview; On The Internet March/April 1998; pp. 39-40

(6) Pimienta, Daniel; Étude langues, cultures et Internet 3ème édition mars 1998: presented to the conference "Visionarios"; Caracas 22-24 April 1998;

(7) TechWeb 9 Mar 98

[INET'98] [ Up ][Prev][Next]