Amazon (AMZN) Web Services (AWS) serves as the backbone of some of the most popular apps and websites in the world. Which is why when the company’s Simple Storage Solution (S3) went down last week, a good chunk of the internet went with it.
The culprit of the outage that caused countless individuals to go a few hours without sending Bitmojis, using their smart doorbells and reading articles totally unrelated to their jobs? A typo.
The good news: Amazon has addressed the issue, and is working to ensure nothing similar happens again. Indeed, the incident last week was an aberration that shouldn’t scare you into not trusting AWS anymore, according to two analysts we spoke to.
The typo that shut down the web
So how did a single typo take down so many websites and services to begin with, though? Well, it turns out an engineer servicing Amazon’s S3 system basically pressed the wrong button and, rather than taking a handful of servers offline for servicing, took a whole slew of them offline. And you thought you had a rough day at work when you had to cover Jenna’s closing shift.
As it turns out, Amazon hadn’t performed a complete restart of its S3 system in years. That, coupled with the fact that the S3 service has grown so quickly, meant getting everything back online took a bit longer than the company expected.
Eventually, the servers came back and the affected apps and websites were up and running normally. So what does all of this say about Amazon? Can companies still trust AWS if a typo can bring parts of it, and the internet, down?
A surprising failure
Actually, yes. In fact, everybody was so surprised about last week’s ill-fated typo because AWS is usually so reliable. Lydia Leong, VP distinguished analyst at the industry research firm Gartner, says Amazon’s AWS S3 availability was 100% in 2016.
There was an outage last year, but that had to do with a large denial of service (DDoS) attack affecting a third-party DNS service provider Amazon uses, not AWS itself.
According to Leong, Amazon actually guarantees 99.9% availability of its AWS service. If it falls below that threshold, the company will reimburse its clients 10% of their fees.
Forrester Research analyst Dave Bartoletti offered similar thoughts on the AWS outage, saying, “I do not think this erodes trust in AWS to a significant degree, since failures in AWS’s S3 service are SO rare.”
Bartoletti said that Amazon clearly learned from the issue — the company even posted a post mortem of the incident online explaining the outage — and that service was restored within hours.
“Overall, S3 was and is one of the most reliable shared storage services in the cloud,” he added.
Amazon’s not alone
Amazon isn’t the only cloud services giant to experience outages. As Bartoletti noted, both Microsoft’s (MSFT) Azure and Google’s (GOOG, GOOGL) cloud offerings have gone offline before. And according to Leong, cloud services provider Joyent suffered an incident similar to Amazon’s in 2014.
Unfortunately, Amazon’s popularity is also a problem onto itself. “The sheer scale of market dominance of AWS shows a broad risk with one provider,” explained Leong. “An outage with AWS tends to impact a disproportionate amount of the internet.”
How much larger is Amazon’s cloud service than its competitors? According to Bartoletti, the company generates 5 to 7 times the revenue from its public cloud than its nearest competitors do. Let’s just hope for the company’s sake, and the sake of our ability to procrastinate while playing with smartphone apps, that Amazon doesn’t experience any further issues in the near future.
More from Dan:
Email Daniel at email@example.com; follow him on Twitter at @DanielHowley.