In part 1, I gave two predictions: that the Ukrainian Internet would prove resilient as Russia’s attacks on infrastructure continue, and that Central Asian connectivity could become more diverse in response. In this post, I’ll look at how conflict could affect the broader Internet, and predict how that may change the conversation around Internet infrastructure centralization.
Prediction #3: Internet Centralization Will Create Geopolitical Risks That Bad Actors Will Exploit
To a casual observer, the history of the digital economy leads from decentralized innovation toward centralized exploitation by bigger and bigger service providers. This is no accident — most Internet users seem perfectly happy getting their services from market-dominant providers without giving infrastructure any thought. But the price of centralization is always increased invisible tail risk: exposure to truly enormous failures whose impacts are far-reaching. When one provider handles the communication of too many, their failures rapidly become everyone’s failures.
Last July, Rogers Communications suffered a multiday outage that left about 25% of Canadians without access to Internet services. Fixed-line telephony, mobile phones, international roaming, 911 emergency services, and point-of-sale terminals all felt the impact. When I wrote about this event, I emphasized the lack of transparency about what had happened, the long wait for Rogers to regain control of their network, and the concentration risks posed by Canada’s telecom oligopoly.
This time, it was a simple failure of engineering: a maintenance window gone bad. But when telecommunications infrastructure reaches a sufficient size that its outages have national-scale impacts, geopolitical concerns like those faced by Ukraine inevitably enter the picture.
In 2023, countries whose Internet economy is insufficiently diverse will run the risk of having their national communications infrastructure held hostage in geopolitical grudge matches. The largest ISPs and mobile providers have now joined the energy distribution networks as an almost irresistible target for those who seek to demoralize civilian populations outside the formal bounds of the law of armed conflict. At some point, infrastructure concentration becomes everyone’s national security concern.
Prediction #4: Extended War Will Lead to Reconsideration of Sanctions-related Internet Disconnections
This observation is a coin with two sides. Once infrastructure denial is in play, its use to enforce international sanctions is in play as well. The Internet community was split in early 2022 over whether sanctions should apply to telecommunications, and whether IXPs should be shutting off peering sessions to avoid servicing sanctioned entities within aggressor countries.
In March, I was among the voices that spoke in favor of maintaining connectivity as the Ukraine-Russia war raged, and I opposed the creation of blocklists that would help coordinate targeted Internet shutdowns as part of enforcing international sanctions. In the end, not much disconnection took place, and the Russian Internet survives pretty much in its pre-invasion state.
Remember, Russia’s Internet ecosystem inherited significant multi-provider diversity from its origins in the chaos after the collapse of the Soviet Union, so it’s fairly resistant to deliberate disconnection. But some of that diversity was rolled back in the consolidation of international transit behind the revitalized Rostelecom. The national provider, plus a few very large mobile providers, now controls the international flow of traffic for much of the country, and their disconnection would certainly have a material impact on the Russian Internet experience.
Will this matter in 2023? If Russia continues to target critical infrastructure, whatever initial squeamishness the Internet operations community may have felt about imposing Internet shutdowns could evaporate, particularly as the 24 February anniversary approaches. That would be a door that would be difficult to walk back through.
Prediction #5: Attention May Turn to Decentralizing Internet Infrastructure
We know that by pushing our engineered systems toward higher diversity, and by incorporating more thoughtful decentralization, we can improve their resilience in the face of natural disasters, human aggression, and simple bad luck. But throughout its history, the Internet seems to have veered from one centralization trap (captured by state-owned PSTN-derivative monopoly ISPs) squarely into others of our choosing (connect-to-everyone social media, host-everything cloud platforms).
The Twitter crisis has driven millions of users to explore Mastodon and the Fediverse, not only fleeing the walled garden of Twitter’s algorithms, but opening new conversations about the proper role of small infrastructure operators in content moderation, identity maintenance, and decentralized community formation.
What’s different here?
Perhaps we’re rediscovering that the infrastructure of digital society isn’t costless or inherently value-neutral. Perhaps we’re remembering that trading away responsibility and complication to centralized infrastructure operators is a deliberate choice that eventually has an associated price.
As Twitter’s infrastructure continues to degrade in 2023, creating visible failures like the one that affected Australia and New Zealand, some of that newfound energy and consumer skepticism may find its way into conversations around the decentralization of the underlying Internet infrastructure as well.
At that point, it will be the job of the Internet measurement community to press the point forward, celebrating regions of the world where the Internet is diverse, decentralized, and growing, and focusing attention on regions with fewer connectivity choices and greater need. By highlighting the failures and successes of infrastructure, we are reminded that while the Internet is for everyone, it also takes everyone’s care and attention to build and maintain the Internet.
Here’s hoping for a productive and more peaceful resolution to 2023.
Jim Cowie has more than 30 years of entrepreneurial and software development experience in large-scale data collection and analysis, high-performance computing, Internet measurement, and data-driven investment strategies.