Waking up to a world without internet might seem liberating, but you may find yourself pondering your next steps.
If you have a checkbook handy, consider using it to purchase some groceries. Should your landline still function, you can reach out to your employer. Then, as long as you still remember how to find your way without modern navigation, a trip to the store is possible.
The recent outage in a Virginia data center highlighted that while the internet is a crucial component of contemporary existence, its foundation rests on aging systems and physical components, leading many to question what it would take for it to come crashing down.
The answer is straightforward: a streak of bad luck, deliberate cyberattacks, or a combination of both. Severe weather events can knock out numerous data centers. Unexpected triggers in AI-generated codes at significant providers like Amazon, Google, and Microsoft could lead to widespread software failures. Armed interventions targeting critical infrastructure could also play a role.
Although these scenarios would be devastating, the more significant concerns for a select group of internet specialists revolve around sudden failures in the outdated protocols that support the entire network. Picture this as a plumbing system that manages connection flows or an address directory that allows machines to locate one another.
We refer to it as “the big one,” but if that occurs, having a checkbook on hand might be crucial.
Something substantial could commence When a tornado swept through Council Bluffs, Iowa, it ravaged a set of low-lying data centers critical to Google’s operations.
This region is known as us-central1, one of Google’s data center clusters, vital for various services including its cloud platform, YouTube, and Gmail (2019) power outages reported here took place that affected users across the United States and Europe.
As YouTube cooking videos become glitchy, dinner preparations go awry. Employees worldwide rush to update emails that suddenly vanish, resorting to face-to-face communication instead. US officials noted a deterioration in certain government services before refocusing their efforts on a new operation against Signal.
While this situation is inconvenient, it doesn’t signify the end of the internet. “Technically, as long as two devices are connected with a router, the Internet functions,” states Michał “Risiek” Wojniak, who works in DNS, the system linked to this week’s outage.
However, “there’s a significant concentration of control happening online,” points out Stephen Murdoch, a computer science professor at University College London. “This mirrors trends in economics: it’s typically more cost-effective to centralize operations.”
But what if extreme heat wipes out US East-1, part of the Virginia facility housing “Data Center Array,” a crucial node for Amazon Web Services (AWS), the epicenter of this week’s outage, as well as nearby regions? Meanwhile, a significant cluster in Europe suffers a cyberattack. frankfurt or London. As a result, the network may redirect traffic to a secondary hub (a less-frequented data center), which subsequently faces capacity issues akin to a congested side road in Los Angeles.
Alternatively, if we shift focus from disaster scenarios to automation risks, increased traffic might unveil hidden bugs within AWS’s internally revised infrastructure, possibly an oversight from months prior. Earlier this summer, two AWS employees were let go amid a broader push towards automation. Faced with an influx of unknown requests, AWS begins to falter.
The signal will falter, and so will Slack, Netflix, and Lloyds Bank. Your Roomba vacuum becomes silent. Smart mattresses may misbehave, just like smart locks.
Without Amazon and Google, the internet would be nearly unrecognizable. Together, AWS, Microsoft, and Google command over 60% of the global cloud services market, making it nearly impossible to quantify the number of services reliant on them.
“However, at its core, the Internet continues to operate,” remarks Doug Madley, an expert in internet infrastructure who studies disruptions. “While the usual activities may be limited, the underlying network remains functional.”
You might believe the biggest risk lies in attacks on undersea cables. While this notion captivates think tanks in Washington, little action has materialized. Undersea cables incur regular damage, Madley notes, with the United Nations estimating between 150 to 200 faults occurring annually.
“To significantly impair communication, a vast amount of data must be disrupted. The undersea cable sector often asserts, ‘We manage these issues routinely.’
After newsletter promotion
Subsequently, a group of anonymous hackers targets a DNS service provider, a key player in the Internet’s directory system. For example, Verisign manages all online domains ending with certain “.com” or “.net” suffixes. Other providers oversee domains like “.biz” and “.us.”
According to Madley, the likelihood of such a provider being taken down is minimal. “If anything were to happen to VeriSign, .com would vanish, which presents a strong financial motivation for them to prevent that.”
To genuinely disrupt the larger ecosystem, a colossal error involving fundamental infrastructure beyond Amazon or Google would be required. Such a scenario would be unprecedented; the closest parallel occurred in 2016 when an attack on Dyn, a small DNS provider, brought down Guardian, X, among others.
If .com were to disappear, essential services like banks, hospitals, and various communication platforms would vanish too. Although some elements of the government’s internet structure remain intact, such as the U.S. secure messaging system Siprnet.
Yet, the internet would persist, at least for niche communities. There are self-hosted blogs, decentralized social networks like Mastodon, and particular domains like “.io” or “.is.”
Murdoch and Madrid contemplate a drastic scenario capable of eliminating the rest. Murdoch alludes to a potential bug in the BIND software supporting DNS. Meanwhile, Madrid emphasizes testimonies from Massachusetts hackers who informed Congress in 1998 about a vulnerability that could “bring the Internet down in 30 minutes.”
This vulnerability pertains to a system one layer above DNS: the Border Gateway Protocol, directing all web traffic. Madley argues that such an event is highly improbable, as it would require a full-scale emergency response, and the protocols are “incredibly resilient; otherwise, we would have already experienced a collapse.”
Even if the internet were to be entirely shut down, it’s uncertain whether it would ever reboot, warns Murdoch. “Once the Internet is active, it doesn’t get turned off. The method of restarting it is not well understood.”
The UK previously had a contingency plan for such a situation. Should the internet ever be disabled, Murdoch notes, individuals knowledgeable about its workings would gather at a pub outside London and brainstorm the next steps.
“I’m not sure if this is still true. This was years ago, and I couldn’t recall the exact pub.”
Source: www.theguardian.com
