A major cloud outage Monday disrupted services across the globe, offering a reminder of how dependent we are on just a few tech giants.
“Amazon Web Services is one of the foundations of the modern internet,” said Shaddi Hasan, assistant professor of computer science at Virginia Tech. “When Amazon breaks, that takes down so many of the services that we use every day.”
Hasan said running a cloud infrastructure is expensive and complex, which is why only a few companies lead the pack.
“There’s only a handful of these companies that run infrastructure at that scale — Amazon, Microsoft, Google and a few others,” he said.
The outage, which began early Monday in Amazon’s Northern Virginia centers, caused widespread disruptions. Services affected ranged from social media platforms to banking apps.
Hasan said “the cloud” is essentially a network of massive data centers or warehouses filled with computers that run the services the public relies on every day. Many of these are located in Northern Virginia, one of the world’s largest data center hubs.
When it comes to cloud computing, there are only a few big players, Hasan said. “Unless you’re in the space, they can kind of be unseen giants to a lot of people. So much of our infrastructure depends on them.”
He said businesses are faced with a tough choice that involves deciding to rely on a major cloud provider and risk being affected by outages, or build and maintain their own infrastructure. The latter is an option that’s often too costly and complex.
Hasan also pointed out that the internet wasn’t always this centralized.
He teaches his students about the early days of the U.S. Advanced Research Projects Agency Network, more popularly known as ARPANET. It was the internet’s predecessor from the 1960s, and was designed to be resilient and even able to work around a nuclear attack.
“But as time has gone on … centralization of the services that run on top of that infrastructure has kind of undermined some of that original ethos and spirit,” he said.
And while outages like this may lead to improvements, Hasan doesn’t expect a major shift.
“It’s hard to imagine a world where … they move away from that model,” he said.
He said incidents like this highlight just how delicate the system has become: “It reminds us of the fragility of relying on just a few large providers.”
And when things go wrong, fixing them isn’t easy.
“These failures are rare, and when they happen, they’re quite complex to remediate,” Hasan said.