Internet Fragmentation: What’s at Stake?

A few months before the COVID-19 pandemic hit the world, the global Internet Governance Forum (IGF) convened in Berlin under the overarching theme, ‘One World. One Net. One Vision’. Referring to this motto, the Secretary-General of the United Nations, António Guterres, remarked that while he was certain that ‘we live in one world’, it was less sure that ‘we live with only one Net’.

Looking back three years later, it is clear that the existing Internet – the global interoperable network of networks – made huge contributions during the pandemic. Despite the surge in demand for connectivity as people’s lives moved more online (OECD 2020; Damas 2020), the Internet’s technical layer remains robust and stable. It has kept connecting people and communities, enabled remote working, and facilitated online communication with loved ones when aeroplanes were grounded and national borders were closed.

When looking ahead, however, can we take this ‘One Net’ for granted, as our ‘One World’ becomes increasingly fragmented with social and geopolitical tensions? If the global interoperable network is indeed under the threat of fragmentation, what is at stake?

Internet ‘fragmentation’

The debate about Internet fragmentation is not new. Two decades ago, Tim Wu raised concerns about the Chinese government ‘pushing hard to divide that global network into a system of Balkanized national networks’. Worries about states and commercial players adopting policies that might break up the Internet into disconnected segments have recurred ever since, with talk of ‘balkanization’, ‘splinternets’, and – most commonly – ‘Internet fragmentation’.

Yet, what does ‘fragmentation’ of the Internet entail more precisely? The term in relation to the Internet is actually semantically confusing because the Internet is not in fact a single network. It is a mesh of various networks and autonomous systems that communicate with each other. In this sense, the Internet has always been fragmented. Milton Mueller speaks in this regard of a ‘unifragged’ Internet that is simultaneously unified and fragmented (2017b: 26).

What brings these networks together into ‘One Net’ is their interoperability. They ‘speak a common language’ in the form of Internet protocols and unique identifiers (IP addresses and domain names). They also use a single Domain Name System (DNS), a database that translates domain names to IP addresses. At its technical layer, the Internet would only ‘fragment’ if it lost its interoperability: for example, if connected devices or autonomous systems were to use incompatible protocols or if other unique identifier arrangements were to compete with the currently prevailing system. As long as the Internet uses the same universally accepted standards, it does not splinter in its technical layer.

Over the years many concerns have been raised about the possible fragmentation of these standards in the technical layer of the Internet. They include worries about an alternative DNS root (Kohler 2022: 20), the lack of acceptance of Internationalized Domain Names (Perarnaud et al. 2022: 16), and the incompatibility of IPv4 and IPv6 address systems (Drake, Cerf and Kleinwächter 2016; Hill 2012). However, to date, no such developments have instigated a fragmentation of the technical layer of the Internet or a major loss in its interoperability.

The Chinese telecom giant Huawei’s proposal in 2019 for a so-called ‘New IP’ has raised particular concern in the fragmentation debate (Hoffmann et al. 2020; Kohler 2022: 37; see also Nanni in this issue). However, this idea has not so far become an imminent threat. The New IP is more an ambitious concept than a plausibility, and it is unclear whether these proposals could obtain wide enough acceptance to replace existing Internet standards.

The Russian government has proposed to create a separate sovereign Internet segment with its own national DNS: the so-called RuNet. Yet the Russian internet segment remains connected to the global network. Moreover, after Russia invaded Ukraine in February 2022, requests were rejected from the Ukrainian government to remove Russian top-level domain names from the DNS root zone, to shut down DNS root servers located in Russia, and to revoke the right in Russia to use IPv4 and IPv6 addresses (Ministry of Digital Transformation of Ukraine 2022a, 2022b). In rejecting the Ukrainian requests, letters from the Internet Corporation for Assigned Names and Numbers (ICANN) and the European IP Networks Network Coordination Centre (RIPE NCC) referred to the need to maintain neutrality, stability, and interoperability of the Internet. Global connectivity is still here and the DNS and Internet protocols remain at its core.

So is fragmentation a myth?

If the Internet remains interoperable in its technical layer, is the debate about fragmentation irrelevant? Au contraire, the discussion is as pertinent today as ever. To understand its significance, we have to go beyond the technical layer and unpack other dimensions of the term ‘fragmentation’.

Many scholars when referring to fragmentation or splintering of the global network discuss processes that happen beyond the Internet’s technical layer to include the ‘information layer’ (Hill 2012), the ‘legal layer’, and the ‘application and content layer’ (DeNardis 2016). Drake, Cerf and Kleinwächter (2016) frame this issue as ‘governmental’ and ‘commercial’ fragmentation. These discussions concern differences in national and corporate approaches to restricting global information flows and divergencies in how Internet users can access, send, and receive information on the Internet, due to policies adopted by governments and business enterprises.

National policies and regulations that limit the exchange of information on the global network can be adopted for various reasons (Haggart et al. 2021). Some are driven by genuine cybersecurity concerns and the protection of users from crime, disinformation, and other threats. However, authoritarian regimes may also introduce restrictions on information flows in order to silence dissent and protect their own survival. The tools to implement such limitations are strikingly similar from one country to the next: data localization laws, various content filtering and blocking practices, censorship, removal of applications, and territorialization of information flows. On the level of telecommunication infrastructure, a government can impose Internet shutdowns via communication providers as an ultimate way to control the Net within its territory (Perarnaud et al. 2022: 24).

The fragmentation debate may also refer to policies, practices, and business models of commercial companies that limit global information flows and diverge users’ experiences (Drake, Cerf and Kleinwächter 2016). Such practices can include companies restricting access to content depending on the user's geographical location and slowing down or limiting internet traffic based on the types of services and applications. There are also concerns about users being locked into corporate ‘walled gardens’ where commercial providers such as digital platforms control customers' access to content and services and govern users via their Terms and Conditions (Perarnaud et al. 2022: 24, Drake, Cerf and Kleinwächter 2016).

Some scholars have objected to taking the notion of Internet fragmentation beyond the technical layer. Goldsmith refers to fragmentation as ‘the scary-sounding but in fact empty notion’ (2019: 822). Mueller (2017a, 2017b) suggests to speak not of ‘fragmentation’ but ‘alignment’, meaning a process whereby states ‘align’ digital information flows with their borders, attempting ‘to force the round peg of global communications into the square hole of territorial states’ (Mueller 2017a). Others speak in this regard of ‘territorialisation’ of the global information network and of ‘Internet conflict’.

No matter which terminology we use, the increasing attempts of national governments to control information flows is alarming. More and more frequently, states treat the Internet as an extension of their sovereign soil (Munn 2020) rather than a global, open, interoperable network. A few years ago, the debate about territorial constraints on global information flows and Internet sovereignty was targeted at authoritarian political regimes. Nowadays, democratic states and the European Union (EU) are also building Internet policies around the concept of ‘digital sovereignty’. Both authoritarian and democratic states alike are trying to exercise more and more control over the global network (Haggart et al. 2021). This ‘fragments’ governance and regulatory approaches, leads to significant divergences in the experience of Internet users, and diminishes the value of ‘One Net’ as a tool for communication, empowerment, and innovation.

One World, One Net, One Vision?

Let’s look back at the IGF 2019, where the UN Secretary-General spoke of living in ‘One World’ but not necessarily ‘One Net’. The actual situation is the opposite. We do have One Net, and this network retains its full potential for global interoperability and connectivity. True, some states seek to limit this potential with policies like content censorship, territorialization of information flows, and data localization. However, the promise of global connectivity in ‘One Net’ persists as long as the network remains interoperable. The core of this interoperability – the IP numbers, the DNS, the data transmission protocols – were developed not by states, but by a global community of nongovernmental actors.

Yet we do not live in ‘One World’. This world is divided by states, territorial borders, and a host of geopolitical, social, and economic tensions. While the world increasingly fragments socially and politically, the technical layer of the Internet has so far remained interoperable, even during crises, emergencies, and wars. Perhaps this stability arises since the technical layer is not in the hands of states, but global multistakeholder structures such as ICANN.

Yet this multistakeholder model is increasingly endangered by states’ appetite for more regulation and control of the Internet. Governments like Russia and China continuously push to transfer control over the technical layer to multilateral bodies such as the International Telecommunication Union (ITU) (Bertuzzi 2022; Scott and Goujard 2022). Even the EU, which declares its continuous support for the multistakeholder governance model, is currently discussing various initiatives that directly impact the nongovernmental governance of the DNS (see further Plexida in this issue). These steps include, for instance, the proposed second EU Network and Information Security Directive (NIS2), which has the potential to circumvent the current multistakeholder processes and impose a patchwork of governmental regulations on the global DNS (Mueller 2020; ICANN 2021; Perarnaud et al. 2022: 40-41). The resulting damage to the current governance model of the DNS would severely impact stakeholder trust in the technical layer of the Internet and its stability, interoperability, and resilience. If we move in this direction, the ‘One Net’ that helped the world get through the global health emergency of Covid-19 might be much less reliable, stable, and robust when the next crisis hits.