Anyone who doubts the extent to which Amazon rules our lives — way beyond the things we buy — need only look at the spiraling effects of an Amazon Web Services (AWS) outage on December 7, when everything from banking apps to home delivery to the Christmas lights suddenly blinked out.
AWS, which provides nearly a third of all cloud computing services used by businesses online, first reported an outage on the morning of Dec. 7; the problem lay in the fact that several network devices were flooded with a lot of traffic from unknown sources. By evening, AWS said on its status dashboard that it was still “working on full recovery for all services.” By the next morning, the main glitches had been resolved, although AWS continued to be slower than usual.
Therefore, for much of the day of December 7, a wide variety of web-based services became lethargic or comatose. Our daily routines have become profoundly but invisibly dependent on cloud computing, so even a partial AWS problem can cause an astonishing suspension of our regular activities. Indeed, a particularly unlucky person could have had any part of their day affected by the outage, bringing normal life to a standstill.
What services did the AWS outage affect?
Our hypothetical consumer, a college student, woke up late, just around the time AWS went on the fritz, so she couldn’t order her voice-activated coffee maker to serve an espresso; Alexa, the Amazon assistant now built into several coffee makers, had gone down. Her fridge froze as she tried and… do not update its operating system. Even her smart light bulbs refused to turn on to dispel the winter gloom.
She had a Zoom call in the morning: a group study thing, organized through the online platform Meetup, but it was canceled because Meetup used AWS at. To pass the time, she thought she would watch some Netflix; no luck there either. She couldn’t order groceries on Instacart, or rebook her Delta Air Lines ticket for a Christmas Eve trip, or send money to her boyfriend through Venmo, or buy meme stocks on Robinhood. She expected an Amazon package, but delivery guys found that their apps weren’t working either, so they didn’t know where to go or what to deliver. Maybe it didn’t matter, because the driver couldn’t have rung her Ring doorbell anyway.
The problems continued. She didn’t have access to Canvas, the popular online education platform, so she didn’t know what to study for her final exam. Tinder didn’t even work. With all this time on her hands, she decided to clean up her apartment for a long time, but discovered that her Roomba vacuum also relied on AWS.
Elsewhere, visitors to Disney theme parks couldn’t check in online or pay for purchases, Adele fans were foiled in their plans to buy tickets for her upcoming tour, webcasts at a UBS conference had to be rescheduled, and crypto geeks couldn’t obsessively check the value of their assets on Coinbase or other exchanges. (At Quartz, we had trouble sending some of our newsletters.) A Twitter employee in San Antonio found that AWS dysfunctions had even affected his Christmas lights.
Given how complex the architecture of the Internet is, outages are surprisingly rare; When servers from Fastly, a cloud computing company, went down in June, it proved a lesson in the virtue of decentralization: the practice of distributing content across as many servers as possible. However, AWS’s outage reminds us that the Internet suffers from other forms of centralization, in this case from an overwhelming reliance on Amazon, which dominates the online services that complement our offline lives.
Correction: An earlier version of this stated that AWS provides cloud computing solutions to nearly a third of the Internet. In fact, it provides nearly a third of all cloud computing services used by businesses online.