San Francisco Frontiers - Internet services cut for hours by Amazon cloud outage

The disruption affected streaming platforms, including Amazon's Prime Video service and Disney+, as well as Perplexity AI, the Fortnite game, Airbnb, Snapchat and Duolingo.

Mobile telephone services and messaging apps Signal and Whatsapp were also affected in Europe, according to Downdetector, a website monitoring internet problems.

Some UK banks, such as Lloyd's were also impacted, and pointed to Amazon Web Services (AWS) as the problem.

AWS handles nearly a third of the planet's cloud infrastructure market, powering millions of apps and websites around the world.

AWS's maintenance site said its engineers scrambled to fix the underlying DNS issue once they became aware at 0711 GMT of "increased error rates" hitting multiple services.

More than four hours later, though some problems persisted, AWS said it was on the path "towards full recovery" and most of its operations were "succeeding normally".

The outage showed "how reliant we all are on the likes of Amazon, as well as Microsoft and Alphabet for many of the online services we more or less take for granted," said financial analyst Michael Hewson.

"On an economic level it's almost akin to putting all of your economic eggs in one basket."

- Cloud leader -

AWS leads the cloud computing market, followed closely by Microsoft Azure, and with Google Cloud in third place. Businesses, government and consumers around the world rely on their infrastructure for online activities.

The British government's websites were among those affected by Monday's outage, according to Downdetector, which relies on users to signal online problems they encounter.

AWS said the problem appeared to originate in one of its critical infrastructure zones, the US-East-1 region, located in the US state of Virginia.

The DNS issue identified has to do with the Domain Name System protocol used on the internet to associate information with various domain names.

Although the DNS problem was "fully mitigated" at 1030 GMT, it caused a huge backlog of stymied requests that had to be worked through, AWS said.

"This major online outage underscores a stark reality: Business operations associated with one critical vendor in a region can cascade into global instability," said independent cyber analyst Rimesh Patel, quoted by the Science Media Centre, a UK charity focused on science and engineering.

"It highlights how interdependent our digital infrastructure has become," said Alan Woodward, a University of Surrey professor quoted by the same centre.

In July 2024, another global online outage occurred when a US cybersecurity company, CrowdStrike, issued a faulty update to its software used by airports, hospitals and many organisations.

According to Microsoft, some 8.5 million devices were affected, resulting in a systems crash and users being confronted with a "blue screen of death".

T.Martin--SFF