It's the morning after the internet went down, and as much as you'd like to think you'll be ecstatic, you're probably wondering what to do.
You can buy groceries using a checkbook if you have one. Call work from a landline – if yours is still connected. After that, you could go to the store, the main thing is that you still know how to navigate without 5G.
A data center failure in the US state of Virginia this week reminded us that the unlikely is not the impossible. The Internet may have become an indispensable backbone of modern life, but it also represents a network of outdated software and physical infrastructure, leaving some wondering what it will take to tear it all down.
The answer could be as simple as acute failure, several targeted attacks, or both. Extreme weather conditions destroyed several key data centers. A line of code written by artificial intelligence deep inside a large vendor such as Amazon, Google or Microsoft – fires unexpectedly and causes a cascading software failure. An armed group or intelligence agency intercepts a pair of submarine cables.
That would be bad. But the real doomsday, which the world's few Internet experts in private Slack groups are still worrying about, is a little different – a sudden, avalanche-like error in the creaky, decade-old protocols that underpin the entire Internet. Think of a pipeline that directs the flow of connections, or address books that allow one machine to find another.
We'll call it the “big thing,” and if it happens, you'll at least need your checkbook.
The big one may begin as a summer tornado tears through Council Bluffs, Iowa, devastating the small cluster of data centers that are integral to Google's offering.
This area, called us-central1, is a cluster of Google data centers critical to its cloud platform, as well as YouTube and Gmail. shutdown Here shot down these services in the US and Europe.
Dinners are on fire and YouTube cooking videos are dying. Workers around the world frantically refresh emails that suddenly become unavailable, then resign themselves to face-to-face communication. Senior US officials notice that some government agencies have slowed down before returning to planning a new attack on Signal.
All this is inconvenient, but the end of the Internet is still far away. “Technically, if we have two network devices and a router between them, the Internet works,” says Michal “rysiek” Wozniak, who works at DNS, the system that suffered the outage this week.
But there is “absolutely more concentration” on the Internet, says Stephen Murdoch, professor of computer science at University College London. “It comes with the economy. It's just cheaper to manage everything in one place.”
But what if, then, a heat wave in the eastern US destroys US East-1, part of the complex in Virginia that houses the “Data Center Alley”, a key hub for Amazon Web Services (AWS) that was the center of an outage this week – among a handful of its neighbors. Meanwhile, a cyber attack hit a large European cluster, they say Frankfurt or London. As a result, networks redirect traffic to secondary nodes, less-used data centers that, like driveways in Los Angeles traffic, quickly become unusable.
Or, if we stray from the disaster movie to the dangers of automation, the increased traffic could cause a bug in the internal AWS infrastructure that was rewritten by artificial intelligence months ago—perhaps one that went undetected. after hundreds AWS employees were laid off this summer as part of the company's broader push toward automation. Overloaded with unfamiliar requests, AWS begins to crash.
The signal disappears. Slack, Netflix and Lloyd's bank do the same. Roomba vacuum cleaners are silent. Smart mattresses are coming scoundrel and smart locks fail.
WITH Amazon and Google, the Internet will look largely unfamiliar. AWS, Microsoft and Google together account for more than 60% of the global cloud services market – and it is almost impossible to estimate approximately how many services depend on them.
“But the Internet, at its most basic level, is still working,” says Doug Madory, an Internet infrastructure expert who studies the outages. “You just can't do anything online that you're used to because it's all published from these data centers.”
You might think the biggest threat is an attack on a submarine cable. This excites Washington think tanks, but otherwise achieves little. According to Madori, submarine cables break regularly – in fact, the UN estimates that between 150 and 200 failures occur each year.
“You're really going to have to take a bunch of stuff out to affect communications. I think the submarine cable industry would tell you, man, we do this all the time.”
after promoting the newsletter
An anonymous hacker group then attacks the DNS service provider, one of the Internet's phone books. For example, Verisign processes every Internet site that ends with a specific “.com” or “.net.” Another Ultranet handles “.biz” and “.us”.
Madori says that it is extremely unlikely that any of them will ever be destroyed. “If something happened to Verisign, the .com domain would disappear. They have a huge financial incentive to make sure that never happens.”
But to truly disrupt the broader ecosystem, it would take a bug of this magnitude that affects more fundamental infrastructure than Amazon and Google. If this happened, it would be unprecedented – the closest analogy would be the situation in 2016. attack on Dyn, a smaller DNS service provider that shot down Guardian, X and others.
Without .com, banks, hospitals, financial services and most communications platforms will disappear. Some government internet infrastructure will still exist, such as the US secure messaging system. Cyprnet.
And – at least for a strange community of experts – the Internet will still exist. After all, there are self-hosted blogs and decentralized social platforms such as Mastodon, and niche domains including “.io” for the British Indian Ocean and “.is” for Iceland.
Murdock and Madory can come up with scenarios that will devastate the others. Murdoch suspects a bug in the BIND software. language which supports DNS. Madori points to readings from a group of Massachusetts hackers who reported to the US Congress in 1998 a vulnerability that could “bring down the Internet in 30 minutes.”
The vulnerability affected a system one layer above DNS: the border gateway protocol that routes all traffic on the Internet. This is extremely unlikely, says Madori – such an event would be a “can-do” scenario, and the protocol is “super robust, otherwise it would have failed by now.”
If the Internet were ever to be completely shut down, it's unclear whether it could be started again, Murdoch says. “No one has turned off the Internet after it was turned on. No one knows exactly how it can be turned on again.”
In the UK there is a non-virtual contingency plan, or at least there was one. If the Internet goes down, people who know how it works will meet in a pub outside London and decide what to do, Murdoch says.
“I don't know if that's still the case. That was quite a few years ago and I was never told what kind of pub it was.”
