Atelier 1: Des infrastructures durables au-delà des évolutions technologiques

Speaker 1 explains that the Internet was built on open standards developed mainly within the IETF, and that this technical foundation is complemented by a multistakeholder governance model carried notably by ICANN. In his view, the Internet’s evolution rests on these two inseparable dimensions: a technical leg to standardize, and a governance leg to adapt policies and support change.

He first argues that standards were developed mainly within the IETF and that they are open and, for the most part, non-proprietary . He then explains that ICANN is based on a multistakeholder governance model capable of keeping pace with technological change . He illustrates this interplay through multilingualism and cybersecurity: IDNs involved technical work at the IETF followed by integration into ICANN policies , and DNSSEC likewise followed this logic, combining RFCs/standards on the technical side with policies on the governance side .

Speaker 1 argues that ICANN’s multistakeholder model is open to everyone, but that it can only reflect the needs of the whole world if all regions actually take part. When some areas, especially African countries, are absent, their specific concerns do not make it into the debates or the policymaking process.

He explains that part of the world participates insufficiently in ICANN and that some regions have little or poor representation there . He specifies that the role of the staff is to facilitate the attendance and participation of these actors; otherwise only part of the world is represented . He adds that when the countries concerned are not present, their interests are not represented, giving the example of DNS resilience issues in sub-Saharan Africa that other countries will not raise on their behalf .

Speaker 1 argues that the Internet of tomorrow must be shaped collectively by governments, the private sector, technical communities, and users. He warns against a passive posture of mere consumption, which leaves power in the hands of private platforms rather than stakeholder communities.

In his conclusion, he states that there is still an opportunity to co-build the Internet of tomorrow through multistakeholder models and that everyone must take part in it . He contrasts this vision with a purely consumerist use of the Internet, recalling that its original spirit was one of community and multistakeholder governance . He links this participation to concrete goals such as tackling the digital divide, multilingualism, and access to culture, stressing that such progress will not come spontaneously from the private sector alone .

Speaker 1 argues that the sustainability of digital infrastructure is not only about large data centers, but also about the efficiency of DNS processing. He explains that better localization of queries and reducing unnecessary queries can limit energy consumption and therefore the system’s carbon footprint.

He indicates that the DNS carbon footprint does exist, even if it is relatively low, and that some query processing models are more energy-intensive than others . He specifies that avoiding the unnecessary routing of DNS queries outside the territory and limiting cascades of queries constitute a concrete avenue for improvement .

Speaker 1 stresses that in Africa, a significant share of DNS queries is poorly handled and leaves the territory unnecessarily without producing any useful result. He presents the localization of certain DNS infrastructure as a way both to improve technical efficiency and to reduce energy and operational costs.

He explains that work carried out by ICANN in various African countries shows that a large number of DNS queries are useless or fail for various reasons . He mentions a study in South Africa showing that Internet service providers manage DNS queries in an unsustainable way . He adds that ICANN has installed two major structures in Africa to localize DNS queries so that they no longer leave the territory, which has generated savings and improved efficiency .

Speaker 1 believes that the State must guarantee a minimum level of continuity of digital services on its territory, even in the event of an international connectivity outage. To that end, he recommends local mechanisms such as network caching and resilience agreements with operators to ensure the availability of essential government services.

He states that a State must guarantee a minimum level of security on its territory and that a total Internet outage can also disrupt domestic services . He cites as a good practice, particularly for islands and landlocked areas, having cache at the network level in order to maintain at least government services on a local loop . He adds that lack of access to public services and information can severely destabilize the administration and trigger unrest, which justifies specific agreements between the State and operators to ensure local resilience .

Speaker 1 believes that major private platforms operate primarily according to a commercial logic and therefore cannot be the sole foundation of countries’ digital presence. Without the development of local open alternatives, such as national domain names and the open Internet, countries become dependent on closed environments that weaken their sovereignty.

He states that in Africa, the growing influence of the GAFAM is no longer limited to submarine cables but now also extends to satellites . He explains that if countries had made greater use of the Internet as it has existed from the beginning, by developing local DNS and the use of ccTLDs such as .ci in Côte d’Ivoire, they would have been less dependent on platforms . He contrasts domain names and the open Internet with the closed environments of platforms, recalling that on Facebook, the user is on Facebook’s servers and subject to Facebook’s rules, not in a local sovereign space . He concludes that users must be educated to become producers of content on their own domain names; otherwise, sovereignty becomes impossible .

Speaker 1 explains that multilingualism on the Internet has advanced thanks to a combination of technical innovation and policy adaptation. Internationalized domain names first required technical work, and were then incorporated into ICANN policies.

He notes that work was carried out at the IETF level on IDNs and that, at the ICANN level, this evolution was reflected in their inclusion in nearly all policies, particularly during the launch of the new gTLDs . He specifies that there is no isolated special IDN policy, but rather a cross-cutting approach integrated into ICANN’s general policies .

Speaker 1 argues that the opening of new domain names and the development of IDNs serve a political and cultural objective: reducing the historical weight of English and American categories in the visible architecture of the Internet. This should enable more languages, scripts, and territories to exist online in a legitimate and functional way.

He explains that new domain names have been added alongside older ones such as .com or .org, with strong meanings such as .paris or .alsace, linked to geographic identities . He states that the opening of the Internet root was driven in particular by multilingualism and by the desire to move beyond an initial architecture heavily shaped by American usage, citing .edu as an example of a historically American space . He adds that IDNs were developed to allow domain names with accents or specific scripts, but that their success also depends on support by software and forms, including for email addresses .

Speaker 1 considers the use of country-code domains to be a lever for digital autonomy and territorial visibility. He believes countries would benefit from building their digital presence on their own domain names rather than relying primarily on closed applications and platforms.

He explains that the development of mobile Internet in Africa has often made the use of domain names less visible because users mainly accessed services through apps . He gives the example of .ci in Côte d’Ivoire, which he presents as a clear anchor point on the global data map, falling under national jurisdiction . He argues that content production should be encouraged on dedicated domain names rather than on platforms, otherwise dependence on major players will be reinforced .

Speaker 1 responds to the question of knowledge preservation by stressing that ICANN does not decide such directions on its own. He explains that resilience or protection responses can only emerge if communities of interest organize themselves, participate, and effectively bring these issues into governance and standard-setting structures.

He specifies that his role within the ICANN staff is not to develop policies, and that infrastructure resilience models cannot be decided unilaterally by him or by the staff . He stresses that these issues must be decided by the global community and that, without mobilization by interested stakeholders, nothing happens automatically . He adds that structures exist to defend open and interoperable models, but they must be actively occupied in order to advance ideas and drive policy development .

The Moderator steps in to request a clear, educational explanation of the technical acronyms so that the debate is accessible to the entire audience. The Moderator also refocuses the discussion on the session’s guiding question: how to maintain and strengthen sustainable infrastructure beyond technological change, including in the face of AI.

He interrupts Speaker 1 to ask for a simple explanation of DNS, the IETF, and other acronyms, explaining that part of the audience may not be familiar with these concepts . He then reframes the session’s theme by stressing the need to maintain, consolidate, and strengthen sustainable infrastructure beyond today’s technological shifts, especially artificial intelligence, and reminds participants of ICANN’s important role in this reflection .

The Moderator emphasizes that submarine cable breaks do not affect only one country but can destabilize an entire subregion. He concludes that critical infrastructure must be designed with robust alternative solutions, especially for countries in the Global South that are more vulnerable to regional failures.

He cites the crisis in Côte d’Ivoire, recalling that several submarine cables were affected and stressing that this type of outage directly raises the issue of critical infrastructure . He notes that in Africa, a localized disruption, for example in Dakar or Côte d’Ivoire, can affect several neighboring countries, mentioning the case of around five or six countries in the subregion being impacted . He then explicitly raises the question of alternatives, including satellite-based ones, in the face of sabotage or submarine cable cuts .

The Moderator links infrastructure sustainability to its environmental impact and public governance. In his view, robustness, maintenance, geopolitics, and ecological cost must be considered together in order to define more sustainable pathways.

After the presentation on cable robustness, he explicitly asks what these infrastructures cost the environment and what governance approaches and technical solutions could make them both more robust and more environmentally friendly . He places this question in a geopolitical context shaped by potential sabotage and maintenance challenges .

The Moderator points out that the growing number of private investments in cables, satellites, and other infrastructure layers may indeed provide backup solutions, but it raises a question of strategic dependence. He asks whether States should themselves provide or guarantee alternatives in order to preserve their freedom of choice and autonomy.

He notes that powerful private actors are now investing in the sea, space, and network infrastructure, and he asks what prospects this creates in terms of alternatives when a country suffers an outage or sabotage . He raises the question of whether these alternatives should be provided by States rather than being left solely to major private actors .

The Moderator acknowledges that many of the Internet’s structuring concepts remain poorly understood, including by professionals, and that a popularization effort is still needed. He suggests that in-depth explanatory work on the history and architecture of the Internet would be useful to clarify these concepts over the long term.

In response to Speaker 4’s remark, he acknowledges that it would almost require retelling the entire history of the Internet, from the Web to the present day, given how much there is to explain . He concludes by emphasizing the importance of popularization work around these issues .

Speaker 2 stresses the vital nature of submarine cables for global communications and their high vulnerability to cuts and incidents. In his view, this dependence warrants public protection policies, because the stakes go far beyond the private interests of operators.

He states that submarine cables are among the most important infrastructures in the world and carry about 90% of global traffic . He also explains that isolated or underserved areas, particularly in the Global South, are especially exposed when no actor truly takes responsibility for this resilience .

Speaker 2 presents an ITU initiative that brought together experts, authorities, and industry representatives to formulate resilience recommendations. He nevertheless stresses that these recommendations will have real effect only if national governments implement them in their legal and operational frameworks.

He explains that an international consultative forum was established within the ITU, bringing together national authorities, maintenance operators, shipowners, and industry players such as Alcatel, as well as Chinese and American actors . He notes that a summit in Abuja approved 40 recommendations to improve resilience, and that a 200-page report was to be approved the following week to set them out in detail . He adds that half of the recommendations are addressed to the ITU or other international organizations, but that most must be implemented by national governments, which remains difficult because the ITU can only issue suggestions .

Speaker 2 argues that cable cuts should be treated as matters of national sovereignty. He rejects the view that they are merely commercial disputes between private actors and shows that public intervention can have a powerful deterrent effect.

He recounts having helped change the laws in Uruguay and Colombia to sharply increase penalties for fishers cutting cables, raising them from around $1,000 to $500,000 . He specifies that after this reform, incidents of severed cables stopped, which he sees as proof of the effectiveness of the deterrent measure . He emphasizes that it was difficult to convince the authorities that the issue fell under national sovereignty rather than a conflict between private companies .

Speaker 2 explains that the central issue is not only the theoretical lifespan of cables, but the practical ability to maintain and repair them quickly. The shortage of specialized vessels, their aging, and the very high costs of mobilization make maintenance difficult and economically unsustainable.

He notes that there is a shortage of vessels for cable maintenance and that the average age of these ships is very high, to the point that they are almost being withdrawn from service . He adds that many of these vessels are now being used for power cables, oil platforms, or oil surveys rather than telecommunications . He concludes that there is a lack of investment in this area . He then gives some rough cost estimates, explaining that sending a vessel from Curaçao to certain areas can cost around 2 million dollars in fuel, and that even in Europe a repair still amounts to several hundred thousand dollars .

Speaker 2 states that the ownership and usage structure of submarine cables has changed profoundly in favor of hyperscalers and major platforms. According to him, telecom operators are no longer the dominant players they once were, which is changing the sector’s balance and the actual control of infrastructure.

He states that 90% of submarine cable owners today are hyperscalers or GAFAM . He cites the example of the North Atlantic route, where 90% of the bandwidth is used by GAFAM . He further explains that telecom companies may appear as formal owners or capacity lessors, but that the real users are now the major digital players .

Speaker 3 notes that the economic value of the digital sector is increasingly being captured by the GAFAMs, while telecom operators are seeing their revenues and margins shrink. Since investment depends on available resources, this trend is reducing their ability to finance national connectivity infrastructure, even though it remains essential.

He explains that in recent years there has been a shift in ownership of connectivity infrastructure toward the GAFAMs, even though telecom operators have historically been the ones building national connectivity . He then argues that the GAFAMs now concentrate the value created in the digital sector, while telecom operators capture less and less of it, directly reducing their capacity to invest in the necessary infrastructure .

Speaker 3 believes that the sustainability of infrastructure now requires reflection on the sector’s overall economic model. If traditional operators can no longer reinject sufficient resources into networks, digital sovereignty and universal connectivity may be compromised in the medium term.

He explicitly raises the question of what will happen in five or ten years if countries are no longer able to reinvest in their national connectivity infrastructure . He concludes that the sector’s business model may need to be rethought, since all the value is being drawn toward the GAFAMs while traditional players lack new resources to sustain investment .

Speaker 4 argues that domain names are not merely technical addressing elements, but also have an intellectual function in organizing information. He compares them to a residual form of cataloguing or a table of contents in a digital environment dominated by search engines and indexes.

He contrasts the former world of knowledge, organized by catalogs, authors, titles, and subjects, with the contemporary Internet, largely structured by indexing and search engines . He argues that, fundamentally, a form of structuring persisted in domain names, giving the example of suffixes such as .biz or .edu, which guide the user as to the type of content or institution likely to be found .

Speaker 4 raises a question of civilizational resilience: how to preserve knowledge, humanist values, and the information heritage if extreme catastrophes were to wipe out the political or cultural structures that sustain them today. He invites us to think about the sustainability of knowledge beyond mere institutional or territorial survival.

Through a metaphor about the possible disappearance of Europe or the Francophonie as physical entities, he asks whether the spirit of knowledge and the encyclopedia could nonetheless continue to exist . He explicitly raises the question of whether there are ways to make the values, knowledge, and heritage circulating on the Internet sustainable so that they survive even the worst catastrophes .

Speaker 4 clearly distinguishes the survival of knowledge from the survival of infrastructures or states themselves. However, he notes that, in today’s world, the Internet has become a major medium for the circulation and transmission of this heritage, which makes the technical question culturally decisive as well.

He states that the survival of encyclopedic and humanist knowledge is a distinct question from the survival of national infrastructures, because nations may disappear without the idea of knowledge necessarily disappearing . At the same time, he emphasizes that the values and heritage we care about now circulate largely in the form of information and heritage on the Internet, directly linking their preservation to the sustainability of digital infrastructures .