Copyright CBS News
The 15-hour disruption to Amazon Web Services, a cloud-based computing service, on Monday shook internet users around the world, disrupting everything from travel to financial transactions — and underscoring the fragility of a system reliant on only a handful of tech giants. Monday's problems at Amazon Web Services, or AWS, spawned 11 million total outage reports, with 3 million reports stemming from U.S. users, Downdetector, a website that tracks online outages, said on Facebook. While other cloud providers support the world's businesses — including Google Cloud Platform (GCP) and Microsoft Azure — AWS has the largest foothold in the market. Since it debuted in 2006, the company has grown to control 38% of the cloud computing infrastructure market, according to research and advisory firm Gartner. Experts say Monday's outage reveals the vulnerabilities of a system where a small number of companies provide such a large chunk of the internet's backbone. "[The] outage is a stark reminder of the massive 'concentration risk' the global economy has accepted by building on a handful of cloud providers," Dave McCarthy, who leads global research for cloud services at the International Data Corporation, told CBS News. AWS' rolodex of customers include government departments, universities and businesses, and platforms such as Venmo, Netflix and Snapchat. "Frankly, many customers may have been unaware that a service they used relied upon AWS and are only learning that now, due to system failures," said Craig Shue, a professor and the head of the computer science department at Worcester Polytechnic Institute. As Shue explained, companies that are not technology-centric may seek out a cloud service like AWS to host their web servers and databases, allowing them to focus on their day-to-day operations. Amazon, meanwhile, handles the more technical aspects such as server maintenance and responses to fluctuations in customer demand, he added. Amazon declined to comment directly on the incident, instead sharing a statement from their website published on Oct. 20. How the AWS outage started The outage started Monday morning in Northern Virginia, home to AWS' oldest and biggest cloud computing hub in the U.S. AWS refers to the cluster of data centers in Virginia as US-EAST-1. The company also has cloud computing hubs in California, Ohio and Oregon. According to expert accounts and Amazon's own explanation, the issue was partially tied to AWS' domain name system (DNS), which is called DynamoDB. As Shue explained, this is the part of the internet that takes domain names like Amazon.com and converts them to IP addresses, numerical labels that identify locations on the internet. The service can be used to distribute traffic to multiple servers to make things faster, Shue said. "It is akin to waiting in a checkout line that then splits off to multiple cashier lanes," Shue explained. "A staff member can direct the customer at the front of the line to the first available server that can handle the job." This function was disrupted on Monday, leading to a backlog of server requests without anything to route them. It's still unclear how the disruption began, experts said. For people on the ground, this meant they were unable to carry out simple web-based functions like checking their flight time on an airline app or sending a payment on Venmo. While the issue started in the eastern U.S. region, its effects were global. Experts say it's difficult to quantify the impact in dollar figures. McCarthy, however, noted that it appears to be one of the most significant outages in recent years. Amazon said that the issue was resolved around 6 p.m. EDT on Monday, or more than 15 hours after the company first reported a problem on the AWS Health Dashboard, where it tracks service disruptions. "This wasn't a minor glitch; it was a cascading failure originating in US-EAST-1, the very nerve center of AWS, which took down everything from gaming and social media to critical financial and educational platforms," he said. Will the outage impact AWS? While the incident may have sparked a small reckoning with cloud services, it's not expected to shake up the landscape too much, experts told CBS News. McCarthy doesn't expect the outage to lead to a mass exodus of AWS customers, but he said it could compel companies to diversify their cloud services so they're not reliant on a single provider. That way if one goes down, the company can still keep some operations going.
