Fastly blamed the global Internet outage on a software bug

Fastly suffered a major glitch this week that briefly took many of the world's most famous websites offline, and the company blamed the problem on a software bug that was triggered when a customer changed Settings. The Fastly problem meant Internet users were unable to connect to a number of popular websites early on Tuesday, including the home pages of The New York Times, The Guardian, Twitch, Reddit and the British government.

In a blog post late on Tuesday, Nick Rockwell, Fastly's senior vice-president of engineering and infrastructure, said: "We experienced a global outage due to an undiscovered software vulnerability that appeared on June 8, when it was triggered by a customer configuration change.

He said the outage was "widespread and serious," but the company quickly identified, isolated and disabled the problem, with most of its network up and running again 49 minutes later. Mr. Rockwell said the flaw was included in a software update that was rolled out in May, and the company was trying to figure out why it had not been detected during testing.

"Although there were specific conditions that triggered the outage, we should have expected it," Rockwell said.

Fastly, based in San Francisco, is a content-delivery network that allows customers to store data such as images and videos on mirrored servers in 26 countries. Getting the data closer to the user means it gets delivered faster.

But the incident highlighted how the global Internet relies on a small number of behind-the-scenes companies such as Fastly that provide vital infrastructure, and it amplified concerns about how they are vulnerable to more serious disruption.