Skip to Content

A total Amazon cloud outage would be the closest thing to the world going offline

Posted on October 24, 2022 by

Categories: AWS

Tags:

One only needs to consider the cascading impacts of an Amazon Web Services (AWS) outage on December 7 to see just how much Amazon controls our lives—far beyond the items we buy. At that time, everything from banking applications to home deliveries to Christmas lights abruptly as cloud went dark.

On December 7, AWS, which provides over a third of all cloud computing services used by businesses online, first reported an outage. The issue was several network devices being overloaded with a significant amount of traffic from unidentified sources. In the evening, AWS stated that it was still “moving towards full recovery across services” on its status dashboard. The significant failures were rectified the following morning. However, cloud  AWS was still operating slower than usual.

Therefore, many web-based services were somnambulant or unconscious for most of December 7. As a result of our complete but invisible reliance on the cloud for our everyday operations, even a small AWS issue can cause an incredible suspension of our regular routines. In fact, a miserable person may have had every single moment of their day affected by the outage, putting a stop to daily life as usual.

What services were impacted by the AWS cloud?

Our fictitious customer, a college student, got up late and couldn’t tell her voice-activated coffee maker to brew an espresso because Alexa, the Amazon assistant now integrated into various coffee machines, was down. Her refrigerator froze up while attempting to upgrade its operating system but failed. Not even her smart lightbulbs would switch on for the winter gloom to be lifted.

She had a Zoom call in the morning regarding a group study activity scheduled through the website Meetup, but it was postponed since Meetup also used AWS. She had planned to watch Netflix pass the time, but that plan also failed. She could not use Instacart to get groceries, rebook her Delta Air Lines flight for a vacation on Christmas Eve, give money to a friend using Venmo, or purchase any stock on Robinhood. She had anticipated receiving an Amazon shipment, but delivery personnel discovered that their applications were not working, leaving them unsure of where to go or what to send. Maybe it didn’t matter because her Ring doorbell couldn’t ring by the driver in the first place.

The issues persisted. She could not access Canvas, the well-known online learning tool, so she was unaware of the material she needed to review for her finals. Even Tinder didn’t function. She decided to finish the long-overdue cleaning of her apartment because she had so much free time, only to find out that her Roomba vacuum machine also used AWS.

Other issues included the ability for Disney theme park guests to check in online or make purchases, the impossibility for Adele fans to purchase tickets for her upcoming tour, the need to reschedule webcasts for a UBS conference, and the inability for crypto enthusiasts to obsessively check the value of their holdings on Coinbase or other exchanges. (At Quartz, we encountered issues with some of our newsletters being out.) An employee of Twitter in San Antonio discovered that AWS issues had even impacted his Christmas lights.

A lesson on the benefits of decentralization—the technique of dispersing material over as many servers as possible—was learned when servers belonging to Fastly, a cloud computing startup, went down in June. Considering the complexity of the internet’s architecture, outages are surprisingly infrequent. But the AWS outage warns that there are other forms of centralization on the internet, including an excessive dependence on Amazon, which controls the online services that adorn our offline lives.