THE IMPERIAL INTERNET:
Middle Eastern Internet Traffic and International Relations
By Kenneth Neil Cukier
The Red Herring
CAINET00 conference; Cairo, Egypt, March 2000.
Internet connectivity today has vital economic and political implications. In the case of the Middle East, international Internet traffic must first pass via a Western country before returning to the region due to the lack of international exchange points. This networking inefficiency albeit justified economically represents one of many new Internet challenges that have foreign policy dimensions, such as access to telecom capacity, and influence on network policies and infrastructure. Developing countries must recognize the new areas of strategic national self-interest inherent in a networked era if they are to protect their sovereignty and economic welfare.
"Westerners [in the 1500s] were putting forth their strength to make themselves masters of the ocean and thereby potential masters of the world. This concentric attack of the modern West upon the Islamic world has inaugurated the present encounter between the two civilizations."
-- Arnold J. Toynbee, "Civilization on Trial," 1948.
"Jobs, knowledge use and economic growth will flow to those societies that are the most connected, with the most networks and the broadest amount of bandwidth."
-- Thomas L. Friedman, "The Lexis and the Olive Tree." 1999.
* * *
I. Introduction: Networks, Old and New.
In the networked world of 1869, new infrastructure was completed to facilitate the flow of traffic and therefore economic activity for the industrial age. That network was of waterways, and a series of interconnected canals, rivers and sea lanes enabled physical goods to be shipped in an early age of globalization. It was, of course, a time when economies were built around tangible goods. And the creation of the Suez Canal in 1869 improved the trade route, reducing travel time to 17 days from Britain to India through the Persian Gulf, rather than the 26 days it took ships to sail around the Cape of Good Hope in Africa. Yet it also marked a new era of geo-politics, as France, and later Britain, stationed troops in Egypt to assure the canals continued operation.
Exactly a century later, in September 1969, an entirely new network was created that would render the earlier network in some respects obsolete by ushering in the information age. In that year, whose 30th year anniversary is being celebrated as these words are written, the first packet-switched data network was created, under the auspices of the United States Department of Defense. It would be the precursor for todays Internet.
And in a manner ironically similar to the Suez Canal, the Internet serves as a new trade route, a modern Silk Road. But one made of glass and light: The fiber-optic cables that enable data to shimmy across continents today transports ideas and digital goods and services as befits an information economy not tangible goods as a century before. However, like yesteryears network of shipping lanes, the digital network despite cyberspaces ephemeral nature also presents geo-political challenges.
This paper examines just one of the new foreign policy challenges that are posed by the emergence of todays network, focusing on how the physical infrastructure of the Internet creates a new economic balance of power among nations. In the age of shipping, time was measured in days and trading partners were based on geography. The Internet does away with that. Network performance is measured in milliseconds, and ones closest trading partner is determined by bandwidth. In some respects, the Internet renders physical geography less important.
However, this is only on the surface. More profoundly, geography does still matter. The Internet is based on a scarce resource, bandwidth, which itself is based on tangible infrastructure: the wires and routers that connect users of different areas in the world. Whether by submarine cable or satellite, the Internet is composed of individual links that stretch from point to point, country to country and continent to continent. Yet all regions of the world do not share the same level of access to the infrastructure on which the Internet runs.
Rather, theres an insidious twist: The countries with the least means to pay for Internet connectivity have to pay the most for it. The developing world, which could benefit greatly from Internet access and need it to remain globally competitive economically, suffer from a scarcity of telecoms capacity. Richer countries in the developed world find that they enjoy the lowest prices for backbone Internet connectivity. This is a historical vestige of the way the global telephone system was built, based on voice-call traffic patterns. Since the overwhelming majority of calls were made among Western countries, telecommunication carriers invested the most in network capacity among those routes. There was a financial incentive to do so, since more calls meant more revenues from that investment.
However, the Internet is data-intensive, requiring more capacity than in the voice-communications world. When Internet usage surged in the mid-1990s, developing countries which already suffered from poor telecom infrastructure found their problems magnified. They lacked the bandwidth necessary to connect to the Internet in a substantial way. This is due to a variety of factors. Firstly, in many cases the infrastructure simply does exist. Either they lack the financial resources to build it, or their economies are not so mature that the amount of telephone use the traditional rationale for building new telecom capacity can justify building it. In some cases, the reasons are due to physical geography. The developing worlds deserts, jungles or sparsely-populated rural populations make building the infrastructure nearly impossible. And in places where infrastructure does exist, the high cost of building it in difficult geographic and political climates means that it is very expensive to use. The situation that is furthered by state-run carriers that face no competition and so have no incentive to drop prices or increase service quality. As a result, the poor state of telecom infrastructure in the developing world remains. The situation holds for national, regional, and intercontinental connections.
II. Internet Economics: Colonial Customers
This places a severe cost on the developing world: If it wants to connect to the Internet, it is forced to effectively become a customer of the West. There are two reasons for this predicament. First is the history of telecommunication economics. The main infrastructure providers come from the developed world, either Europe or the United States, since companies in those markets had the financial means, commercial incentive and technical expertise to build the telecoms infrastructure on which the Internet would later depend. The second reason is due to the way that traffic flows on the Internet. Based on user preferences, the technical composition of the network and economic factors, the Internets data flow finds all circuits lead to the United States, akin to the way all roads led to Rome in an earlier era of imperialism.
Traffic from around the world often first passes via the United States before travelling to its final destination, which in some cases might even be back to the region, or country, it came from. User preferences play a role first in making the U.S. the most heavily trafficked destination because a majority of the content on the Internet is based there. Even content that is created outside the U.S. is often stored on servers in the U.S. because it is less expensive to host it from the country, and also since the best network performance is on U.S.-bound links (thus perpetuating the dominance of the U.S. as a central Internet hub). The underlying technology of the Internet is also a factor. The Internet operates as a network of interconnected networks, and relies on whats called a "routing table" to organize where to send traffic, as it goes from one provider to another. The largest portion of downstream and upstream routes are on U.S. networks, so there is a technical incentive to link directly to the U.S. Internet backbone.
Finally, there is an economic incentive to hub traffic via the United States or another Western country. Since it would be impossible for an Internet service provider in one country to have direct connections with ISPs in all other countries to reach the users there, it is more cost effective to maintain only one large link to the U.S., or a few large links to the U.S. and another Western country, where the ISP can interconnect with other ISPs from around the world. In the case of the Middle East, there are no international exchange points where ISPs from different countries can interconnect to swap traffic on a regional basis. As a result, the U.S. has become the worlds central switching station for Internet traffic. And since U.S.-bound telecom infrastructure is often built and sold by U.S. carriers (originally so that their American customers could telephone others around the world) that bandwidth dedicated to Internet traffic is leased by U.S. companies, in this case, to ISPs in the developing world. Moreover, even in the case where the national telecom carrier of a developing country owns a portion of the capacity itself, the carrier still must pay an interconnection fee to a U.S. carrier to connect to the U.S. Internet backbone and reach the valuable routing table.
That situation is changing, albeit slowly. More regional content from outside the U.S. is remaining local as better infrastructure is built, as exemplified by the situation in Europe after the telecoms sector was liberalized in 1998. Also, a new push by commercial satellite communications providers has offered an expensive yet in some cases cost-effective alternative to submarine cable systems. They sell excess capacity to ISPs outside the US who can link to any another point in the area that the satellite covers, such as Europe. (Some countries in remote parts of the world have always relied on satellite as the international gateway for telephone service, and later Internet service, since land- and sea-based infrastructure does not exist.) Satellite service has enabled some countries in the Middle East and Africa to link to Europe, rather than the U.S., provided it makes economic sense because their end-users predominantly seek content based in Europe. In many cases, this is due to a historical relationship with the European country, often colonial.
Yet the Internet is creating a new form of colonialism. The Internet is by nature hierarchical; all networks are not equal. While all need to interconnect with each other, some networks carry the traffic longer to reach its destination. Also, they have a greater amount of customers and bandwidth, which places them in the position of doing so. That, in turn, gives them dominance. Smaller networks that lack global infrastructure to haul traffic to all destinations are forced to rely on so-called "upstream" network providers to do so, and at a cost. On the Internet, who you are is who you pay. Unsurprisingly, ISPs in the developing world are placed in a position where they must pay upstream networks in the developed world, where there is the best telecom infrastructure, if they are to have Internet connectivity.
III. Network Traffic: Follow the Bits.
This U.S.-centric Internet phenomenon, and its implications, are difficult to quantify. There is a lack of public information about the amount of international bandwidth that passes from one country to another, where it leads, and the company that sold the capacity. However, there are two ways to illustrate the magnitude of the phenomenon.
The first is a networking metric called a traceroute. As its name suggests, it is a means to follow the path that a bit of data takes from one point to another. Tracing the route that Internet traffic takes from the Middle East to another point in world shows whether the data must first pass a third country. It thus reveals how network providers from these third countries stand to profit from the poor lack of telecom infrastructure in other regions.
The second way to demonstrate the issue is to compare the traffic patterns of voice communications with Internet bandwidth connectivity. Specifically, it means examining the percentage of voice calls that are destined to other countries, relative to where the Internet links go to. While Internet users traffic might not coincide with the places that voice callers wish to call, there is a good reason to believe that the differences should not vary much; the Internet is a means of communications like the telephone network, and international communications is often based on social and economic rapports between countries, and especially language.
In a traceroute performed in December 1999 of the actual Silk Road from antiquity, spanning from China to todays Uzbekistan, Internet traffic from one point on the route to another doesnt go directly at all. Instead, data between Luoyang, China and Bukhara in Uzbekistan first passes through Los Angeles and New York, where it then travels to Oslo before eventually making its way to China, likely via satellite. (see Appendix A for the actual table). The traceroute illustrates how Internet traffic takes the least expensive route, which is often not the most geographically direct one. It also shows that the sites ISPs get their backbone connectivity from Western telecom carriers -- in this case, Sprint and Teleglobe -- who hub their traffic in the U.S. and Europe. (While this demonstrates the point for the sake of this paper, the author intends to show more recent traceroutes from Egypt to other destinations in the Middle East when the paper is presented publicly during the CAINET00 conference.)
The relationship between telephone traffic and Internet bandwidth tells the same story, but from a macro perspective. Voice call pattern statistics are maintained by national telecom ministries and regulators around the world, which used the data to determine the settlement charges for the international accounting rate regime that requires call-origination networks to pay a fee to call-receiver networks for terminating the call over their infrastructure. Conversely, there are no formal reporting requirements to monitor Internet traffic, since the Internet is a private data networking system. Hence, data on Egypts international Internet connectivity, measured by telecom circuit capacity, is tougher to measure. It must be assembled piecemeal from network operators who may be reluctant to share that data for competitive reasons. Also, due to the fast-pace of Internet deployment, the amount likely changes often.
A comparison of international voice calls and Internet bandwidth shows that Egyptian Internet traffic must transit a third country to reach users in other areas, in some cases back in the Middle East. In 1998, the most recent statistics available, Egypts top calling destination was Saudi Arabia, which accounted for 15.7% of all outgoing international calls from the country that year (see Appendix B for the full table). It is followed by the U.S. (8.5%) and the United Kingdom (8.4%). Eleven Middle Eastern countries appear in the top 20 calling destinations, and account for 55.5% of outgoing traffic. The true figure for all calls within the region is larger countries outside the top 20 account for 21.1% of Egypts outgoing calls, of which some can be presumed to be in the Middle East.
However, in regards to Internet traffic there were no direct links between Egypt and any other country in the region in 1998, the time when the most recent statistics were collected by TeleGeography Inc., a Washington, DC-based telecom statistical research firm. According to the findings, certainly out of date today but still useful to explain the situation, in 1998 Egypt had only two Internet connections outside the country, with a total capacity of 3.5 megabits-per-second to the U.S. and 1.5 megabits to France. A separate 1998 study by IDATE, a telecom market research firm based in Montpellier, France, corroborated TeleGeographys findings, yet noted that an association called EGTI had a direct connection with Italy on the SEA-BONE circuit (IDATE did not record the amount of capacity on the link).
Taken together, the call and Internet traffic figures show that while the majority of Egyptian telephone communications remain within the Middle Eastern region, all of its Internet traffic that travels either to or from other regional countries first passes via a third country, in the West. These findings suggest that Egypts Internet architecture specifically, and the Middle Easts Internet more generally, is less efficient than it might otherwise be were the traffic able to travel via the most direct path. The implications are that Egyptian users ultimately pay Western network operators for access to the global network.
The main reason for this state of affairs is the paucity of telecommunications capacity region-wide. Other than expensive and poorly-performing satellite connections, there are only three submarine cable systems that pass through the Middle East: SEA-ME-WEA 2 (built in 1994 with 1.12 gigabits-per-second of capacity); FLAG (built in 1997 with 10 gigabits of capacity); and SEA-ME-WEA 3 (built in 1999 with 80 gigabits of capacity). The SEA-ME-WEA systems get their name because the cables pass from South East Asia through the Middle East and on to Western Europe and America. Encouragingly, this situation is set to improve slightly: In December 1999 Telecom Italia announced plans to deploy a 7,000 kilometer fiber-optic subsea cable in the Mediterranean, mainly for Internet traffic, connecting Italy, Egypt, Greece, Turkey and Israel.
IV. Conclusion: New Foreign Policy Considerations
"If the Suez Canal is our back door to the East, it is the front door to Europe of Australia, New Zealand and India. It is, in fact, the swing door of the British Empire, which has got to keep continually revolving."
-- Anthony Eden, House of Commons, 23 December 1929.
"A great nations foreign policy involves power, money, trade, oil and arms, but it proceeds from ideas."
-- William Pfaff, "Barbarian Sentiments," 1989.
In 1956, the Suez Canal was the source of international conflict. If its creation at the end of the 19th century was the hallmark of that eras optimism in the existing world order and the basis for a new geopolitical map, then the act by President Gamal Abdel Nassers troops to occupy the canal and expel the foreign-owned holding company marked more than just the collapse of European domination of the canal zone, it sparked the start of a new era in world politics.
Of course, historical events move quicker than societys ability to comprehend the transformation, and traces of the old order were still in operation. The canal, at the time, remained a potent factor in global economics. Traffic through the canal in 1951 was 80 million tons, of which 33% was British. By 1955 that figure grew to 115 million tons of which British cargo represented 28%. And the importance of those loads was immense: In that year, 67 million metric tons of oil and petroleum products passed northward through the canal, which amounted to 76% of the total traffic for the year. And 20.5% of that oil went to the United Kingdom. The canal, in classic foreign policy terms, was a matter of vital national interest for the U.K.
Were one to apply the same sort of analysis today to the Internet, it might seem preposterous. It difficult to conjure up the sentiment that Yahoos commercial interests can be a question of national security. Nevertheless, most experts agree that the Internet is similarly becoming a force in global economics. And a body of similar statistics is emerging for todays networked economy. Analysts claim that by this year, as much as a trillion dollars of electronic commerce revenue will travel over the digital network; yet at the same time, the top ten most popular Internet sites in European nations are dominated by American Web firms, according to the research firm Media Metrix.
As major firms migrate their operations to the Internet, and new ways of measuring the impact of the Internet on the global economy evolves, less apparent are the new foreign policy challenges posed by todays scarce resource Internet connectivity for an age when nations are stitched together with wires, not waterways.
Developing countries must recognize the new areas of strategic national self-interest inherent in a networked era if they are to protect their sovereignty and economic welfare. The first is access to telecommunications infrastructure. Building communication links within the region, and globally, must become a national priority, and organized on a pan-regional basis. Secondly, better network cooperation is needed, especially to construct national and regional exchange points were ISPs can swap local and regional traffic. If better communications capacity existed among nations in the developing world, more regional traffic could remain local and not have to pass via third countries, because the economic incentive to do so would not exist. Together these actions will eliminate the reliance of developing countries on the network providers in the U.S. and Europe. It would replace the current customer-supplier relationship that echoes of the former colonial rapport, by placing the developing world on more equal footing. Meanwhile, as national telecom regulators open their markets to competition, they should be vigilant to ensure that the terms and conditions for networks outside the U.S. to interconnect with U.S. Internet operators do not act as anti-competitive market barriers that are prohibited under the 1997 World Trade Organizations telecom services agreement.
Additionally, developing countries must play an active role in the policy-making institutions of Internet governance, such as ICANN, the Internet Corporation for Assigned Names and Numbers, which sets policy on domain names and Internet Protocol address allocations that are needed for routing. In an earlier era, management of resources like shipping lanes, and later telecommunications, was a matter of formal diplomacy among nations. Today, ICANN requires a similar degree of attention from policy-makers. Although run as private corporation and comprised of non-governmental personnel, it is essentially foreign policy by other means.
The developing world must be involved. Failure to do so, for example, might mean that users in those countries find themselves bound by Western legal traditions that would place a larger cost on them to settle disputes. Also, access to IP addresses are imperative to bringing a country online, and so participation in the ICANN process is necessary. Developing countries need the voice of national government officials to assure that ICANNs policies do not unintentionally but implicitly discriminate against users in their countries. There is a real risk of this happening. Because ICANN is comprised of the private sector, which is inherently strongest in the developed world, its rules could very well end up reflecting their interests at the expense of developing worlds concerns. The developing worlds interests include reasonably-priced Internet resources like domain names and dispute resolution procedures, as well as inexpensive and unencumbered access to IP address space.
These are simple recommendations, yet putting them into place will be extremely difficult. Part of the reason is that there is a strong suspicion among leaders in the developing world that the Internet is simply a new form of Western domination, a new mechanism with which to turn their citizenry into high-tech customers for Western goods. These concerns are not ill-founded. The Internet risks being just that unless the developing world can become producers of content, not just consumers. This entails more than bringing Internet access to the population. It means encouraging the private sector to adopt the technologies to become competitive with Western companies, and participate in the global market place. Also, developing countries should create the legal and financial framework that supports the private sector to develop business models based on a network economy, with which it can attract a global clientele.
The U.S. experience, even stripped of the hubris of an irrational stock market, has shown that the new motor of economic growth and job creation are from small, start-up companies. The Internets technical architecture actually encourages this, by placing power at autonomous end-points with little centralized control. And this plays precisely to the advantage of developing countries, whose businesses necessarily must start small relative to Western firms.
While the U.S.-centric network phenomenon is a discouraging form of network dominance by the West on the developing world, the situation is ultimately an optimistic one. As in the case of canal, the Internet is a two-way network bringing potential benefits to all the parties it connects. It is also neutral in regards to the content that flows over it. It is not neutral in terms of who owns it, but that was quite similar to the case of the Suez Canal, too. If developing countries are to benefit from the revolution brought about as the world moves from a tangible transportation network to another of bits and bytes, they must recognize that control over Internet resources to a large degree influences the control over their own future. It is, as always, their own.
Appendix A: Traceroute of the Silk Road from China to Uzbekistan
Traceroute towards IP-address www.luoyangnanfeng.com.cn from vakt.deponet.uz.
Taken in December 1999, by Sam Paltridge of the Organization for Economic Cooperation and Development (reproduced with permission).
1 * * 4fe0x0-rtr1-r-1-hazz.cn.net (22.214.171.124)
ª 2 126.96.36.199 (188.8.131.52)
ª 3 184.108.40.206 (220.127.116.11)
ª 4 18.104.22.168 (22.214.171.124) 639.971
ª 5 126.96.36.199 (188.8.131.52)
ª 6 * * 184.108.40.206 (220.127.116.11)
ª 7 * 18.104.22.168 (22.214.171.124)
ª 8 * * *
ª 9 * 126.96.36.199 (188.8.131.52)
ª 10 * sl-bb12-ana-9-0.sprintlink.net (184.108.40.206)
ª 11 sl-gw10-ana-8-0-0.sprintlink.net (220.127.116.11)
ª 12 sl-gw10-ana-8-0-0.sprintlink.net (18.104.22.168)
ª 13 sl-teleglob-4-0-0-T3.sprintlink.net (22.214.171.124)
ª 14 if-1-1.core1.LosAngeles.Teleglobe.net (126.96.36.199)
ª 15 if-6-2.core1.NewYork.Teleglobe.net (188.8.131.52)
ª 16 if-5-0-0.bb6.NewYork.Teleglobe.net (184.108.40.206)
ª 18 NO-NIT-TN-1.taide.net (220.127.116.11)
ª 19 * 18.104.22.168 (22.214.171.124)
ª 20 r33-1.uznet.net (126.96.36.199)
ª 21 depositr.uznet.net (188.8.131.52)
ª 22 vakt.deponet.uz (184.108.40.206)
Appendix B: Egypts Largest Telecommunications Routes, 1998.
Destination Minutes (in millions) Percentage of Outgoing Traffic.
1. Saudi Arabia 20.0 15.7%
2. United States 10.8 8.5%
3. United Kingdom 10.7 8.4%
4. United Arab Emirates 8.4 6.6%
5. Italy 8.4 6.6%
6. France 7.5 5.9%
7. Germany 7.3 5.7%
8. Kuwait 7.1 5.6%
9. Lebanon 4.0 3.1%
10. Jordan 3.3 2.6%
Other Middle Eastern countries from top 10 to 20 -- 12.7%
Other countries from top 10 to 20 -- 8.3%
Other destinations -- 21.1%
Source: TeleGeography 1999.
Andrew Boyd. "An Atlas of World Affairs." Methuen & Co. Ltd. London 1957.
K. N. Cukier. "Bandwidth Colonialism?: The Implications of Internet Infrastructure on International E-Commerce." INET'99, San Jose. <http://www.isoc.org/inet99/proceedings/1e/1e_2.htm>
K. N. Cukier. "Rich Man, Poor Man: The Geopolitics of Internet Policy Making." INET'98, Geneva. <http://www.isoc.org/inet98/proceedings/5a/5a_2.htm>
Mark Dessauer and Magda M. Ismail. "E-Commerce in the Developing World: Plug and Play?" OnTheInternet magazine; The Internet Society; Herndon, Virginia; November/December 1999.
Mohamed A. El-Nawawy and Magda M. Ismail. "Overcoming Deterrents and Impediments to Electronic Commerce in Light of Globalization: The Case of Egypt." INET99, Geneva. <http://www.isoc.org/inet99/proceedings/1g/1g_3.htm>
Thomas Friedman, "The Lexis and the Olive Tree: Understanding Globalization." Farrar Straus Giroux, New York, 1999.
Institut de l'Audiovisuel et des Telecommunications en Europe (IDATE). "World Atlas of the Internet; vol. 1." Montpellier, France; 1998. (Principle authors: Khaled Zourray, Roland Montagne, and Boris Saulnier.)
International Telecommunication Union. "Challenges to the Network: The Internet for Development." Second Edition. Geneva, 1999. (Principle author: Ben A. Petrazzini.)
Keith Kyle. "Suez." Butler & Tanner Ltd. London, 1991.
Media Metrix, "Worldwide Internet Measurement Results." 4 February 2000. <http://www.mediametrix.com/PressRoom/Press_Releases/02_04_00.html>
Hans J. Morgenthau, "Politics Among Nations" (Third Edition). Alfred A. Knopf, New York, 1961.
Organization for Economic Cooperation and Development. "Internet Traffic Exchange: Developments and Policy." Paris, 1998. (Author: Sam Paltridge)
William Pfaff, "Barbarian Sentiments: How the American Century Ends." Hill and Wang, New York, 1989.
Gregory C. Staple (ed). "TeleGeography 2000." TeleGeography Inc., Washington, DC. 1999.
Arnold J. Toynbee, "Civilization on Trial," Oxford University Press, London, 1948.
Ernest J. Wilson, III; "Organizing Foreign Policy: A Pragmatic Approach to the Information Revolution." IMP Magazine on Information Impacts; Center for Information Strategy and Policy, of Science Applications International Corp. (SAIC). May 1999. <http://www.cisp.org/imp/may_99/wilson/05_99wilson.htm>
My thanks to Bram Dov Abramson of TeleGeography for providing sound advice for a statistical footing and to Sam Paltridge of the OECD, for first cutting the trail that I respectfully follow here. Magda Ismail of the Egyptian chapter of the Internet Society provided much patience and encouragement as I prepared this work for which I am grateful, as well as for the opportunity to learn from her own research. Lastly, appreciation to my employer, The Red Herring magazine, for the opportunity to report on these themes.
Kenneth Neil Cukier (firstname.lastname@example.org) is the International Editor based in London for the Red Herring magazine, covering the private equity market for high-tech companies, as well as the global political, economic and technical implications of the Internet.
Copyright 2000 Kenneth Neil Cukier - This paper may be freely copied and distributed in portions or in whole, provided that the title and author is identified.