Services from the BBC, The New York Times and the online store Amazon, among others, crashed for an hour on Tuesday.
Several major online services and websites crashed on Tuesday at about 1 pm Finnish time. Among others, the online store Amazon, the social website Reddit, and major international news sites such as the BBC, CNN, and The New York Times ceased operations. The pages slowly returned to operation later in the afternoon.
The cause of the problems was found in the cloud service Fastly, which reported a global disruption on its website. The company announced that it had problems with the global content distribution network and that it was Finnish time in the afternoon to rectify the matter.
“We identified a service setting that triggered a disruption in our POPs worldwide [läsnäolopisteissä], and we have turned that setting off, ”the company said in a statement BBC’s by.
What exactly is the question?
“The idea is to spread servers around the world instead of centralizing them at one point.”
Fastly is a US cloud service that helps other online services and websites bring content and data as close as possible to consumers, says Professor of Computer Networking Jukka Manner From Aalto University.
In practice, with the help of Fastly and other similar companies, websites and various online services work faster for users.
“The idea is to spread the servers around the world instead of centralizing them at one point,” Manner says in a phone interview.
“For example, Amazon can use Fastly to keep some of its data, such as product descriptions and promotional videos, closer to consumers. That way, they load faster and people buy more from the Amazon online store. ”
Tuesday There was a problem with Fastly’s content sharing service. As a result, the sites crashed or stopped working properly for those who had purchased cloud services from Fastly to run more smoothly for their users.
The Fastly bulletin referred to problems with configuration, or settings. The mainland says this points to human error. At least until Tuesday afternoon, Fastly did not say there was a cyber attack in the background.
“Usually when configuration changes are made, they are tested in a test environment. However, there can always be a human error, ”Manner says.
Such crashes can happen without even being noticed in the media. On Tuesday, however, a setting error affected the operations of large media houses as well as other major online services such as Paypal, Twitter, Vimeo and HBO Max. Therefore, it received significant attention for its part.