Rephrasing "Internet fragmentation" (and "the techlash")
Back in 2019, I was concerned primarily with “Internet fragmentation”—roughly, the idea that the Internet is becoming increasingly different in different places.
A fragmented Internet carries consequences for free speech, international business, trade, etc. The topic has businesses, policymakers, and the military spooked. And it hooks into larger debates about “decoupling,” trade protectionism, and concerns about liberalism under threat globally.
A big problem, though: we didn’t—and still don’t—have good theories that explain why Internet fragmentation is happening in the specific ways it is.
Consider the following cases:
The EU’s GDPR, a data protection framework that renders the Internet a bit different in Europe than in the US.
China’s “great firewall,” which censors much of the western Internet (for example, Google).
Russia’s data sovereignty laws, which mandate data about Russians be housed in Russian servers, culminating in a drill that disconnected Russia from the global Internet.
Why did these different actors employ these particular tactics? Why is Russia’s strategy so different from China’s? In general, why is Internet fragmentation so diverse? Without answers to these questions, it’s difficult to imagine how regulators worldwide will react to a changing Internet and a changing geopolitical landscape. That, in turn, makes it difficult to imagine how the Internet might develop.
My argument: the frame of “Internet fragmentation” itself is limiting. I propose an alternative: Internet conflict. I’ll show how this term helps us think more broadly and more specifically about the various clashes we see on the Internet—including some that appear to be distinct from “Internet fragmentation,” but are in fact deeply connected to it.
“Internet fragmentation” buries the lede.
When I started this work in 2019, there were anecdotes about Internet fragmentation, but not much measurement. Is the Internet getting more or less fragmented over time? How and where are these changes happening? I started to measure Internet fragmentation, trying to get a sense of the direction and velocity of Internet fragmentation over time.
Somewhere along the way, a different story emerged. Contrary to the perception that a global Internet is breaking apart, my data shows an Internet growing so highly centralized in the United States that the US could effectively turn the whole thing off. This Internet isn’t “global” in the sense of “global power-sharing.” It’s “global” in the sense of “global hegemony.”What we call “Internet fragmentation” is what happens when other countries say “no.”
From Internet fragmentation to Internet conflict
As Frederick Douzet wrote in 2014, the Internet both reflects and shapes geopolitics. States (and others) regulate the Internet to affect their geopolitical posture; they also take geopolitical action (trade, military alliances, etc) in reaction to shifts on the Internet.The frame of “Internet fragmentation” buries this complexity.
Here’s a frame broader than “the splinternet:”
The Internet is itself a domain of conflict, like geographic territory. Actors compete for power in this domain by controlling, directly or indirectly, physical properties of the network.
Various actors (states and non-states) are opportunistically contesting the US’s dominance to achieve particular domestic and international goals.What we call Internet fragmentation is the observable effect of that contestation.
In other words, Internet fragmentation only makes sense in the context of the US’s dominance over the network. The “splinternet” isn’t a rejection of some global Internet—it’s a rejection of a US-led Internet. And this rejection is more strategic than ideological. It’s an attempt to seize power in a domain when the current hegemon is looking weak.
The Internet conflict framework helps explain why particular countries (and corporations) take particular actions—it helps us see what they might hope to achieve. Here are three cases where this framework gives us something useful.
1. Internet fragmentation vs. the techlash
On one hand, we have “Internet fragmentation.” On the other, we have “the techlash:” emerging efforts to regulate or break up “Big Tech” in the US and abroad. Two different conflicts, right? Well, I’d argue that these two conflicts are deeply connected. “Internet fragmentation” is to Internet conflict between states as the “techlash” is to Internet conflict between states and multinational corporations. Both describe power struggles ‘in’ the Internet, a domain of conflict. And, together, they paint the US as a hegemonic power being actively challenged by both state and non-state actors.
2. Russia’s Internet sovereignty
Why is Russia disconnecting itself from the global Internet? What are they hoping to achieve? Here’s one suggestion I haven’t heard anyone else say: they might want to attack a US-based caching service (imagine a Stuxnet-level attack on Cloudflare). That would bring down the Internet for everyone, but not for them, thanks to their data locality laws.
To me, the Internet conflict framework makes this threat clear, almost obvious. Phrased differently: Internet disconnection drills are consistent with a strategy of increasing escalation of Russian cyberattacks on the US. It also helps us clarify what we can do to ameliorate this kind of attack.
3. A global—but globally censored—Internet
This frame helps us understand the present, but it also helps us imagine possible futures.
Debates around “Internet fragmentation” envision a web of “splintered” internets, each in self-contained national and/or corporate siloes. That vision eludes a highly likely, though infrequently discussed, scenario: an Internet that’s completely global but globally censored.
Imagine an “American Great Firewall” (a “Great Border Wall?”) that censors content globally. This scenario is an “offensive realist” reading of the Internet conflict frame: every actor seeks total domination and will do what they deem necessary to make their domination continue indefinitely.
Internet conflict lets us see struggles over the Internet in a more comprehensive light. It also makes clear that the current hegemon—the US—is fighting a war along multiple fronts, against its allies (the EU), against its adversaries (Russia, China), and against corporations (Google, Facebook, Amazon, Cloudflare).
If a US-dominated Internet is under threat, what Internet comes next? A global Internet led by a consortium of states? A global Internet led by China? By a corporation? Or will we have a multiplicity of Internets with partially overlapping geographies? Will the ensuing conflict be multilateral and purposeful, or power-based and ad-hoc? And will it ever arrive at a stable configuration, or will it devolve into eternal conflict on a barren and unusable web?
To even begin to answer these questions, we need better theories and better data. I have no interest in proposing new terms to be pedantic. Words give us analytical purchase, they make possibilities rise to meet us. They can sharpen our critiques of existing systems and fuel our imagination for building alternatives (technologies, policies, institutions). And they help us understand how well those alternatives are working.
We collect data about Internet conflict, including centralization, continuously using some open-source software we developed. Our data, and a dashboard to view it, is available upon request. You can reply to this email if you’d like to see it.
This work was funded in part by the Internet Society.
This begs the question: if the US has so much power, why is the Internet as open as it is? One explanation—one I'm not sure I believe, but which I haven't heard anyone else propose—is that the Internet may be open because a hegemon dominates it. This runs contrary to the received wisdom that the Internet is most stable when it’s decentralized.
My intention in proposing this heterodox explanation is to show how little we understand about how actors compete for power on the Internet. We need better theories to understand what makes the Internet robust and global (in the sense of global power-sharing).
We observed this correlation between geopolitics and the Internet in our First Monday paper on website blocking.
In the language of relational power theory (power as a relationship in which the behavior of actor A at least partially causes a change in the behavior of actor B), the US has a large domain (there are many B’s) and scope (it can influence many aspects of B’s behavior).
If you want evidence that the US hegemonically controls the Internet, I’ve got lots—lots. But here’s some fresh data to support the reading that Internet fragmentation is a rejection of US dominance: when a government blocks a website, overwhelmingly, that website is based in the US. In the past seven days, 81.9% of confirmed blocking events in the whole world have been against US websites.
This data comes from OONI (Open Observatory of Network Interference). OONI’s data, constantly updated by volunteers worldwide, has people testing specific URLs (inputs) and recording the responses they get back. Through these responses, OONI detects whether there was network interference along the path (most typically, website blocking by nation-states).
We collect both "confirmed" blocking events and "anomalous" events. For each event, we label where the observation was collected and where the website came from. (See more about how we determine where a website is based).
With this data, we make directed pairs; for example, RU->US would capture that Russia interfered with a US-based website. (One way to think about this relationship: the government of X disapproves of behavior that's condoned in Y).
By counting those pairs, we detect how frequently any one country blocks content from any other country. In aggregate, we can detect the direction and velocity of network interference worldwide.