select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
An issue related to Amazon’s web service platform has knocked out Snapchat, Duolingo, and a number of other websites, apps and games. Online banking services in the UK are also affected. The All Ireland Science Media Centre asked local experts to comment.
Dr Domhnall Carlin, Research Software Engineering Fellow, School of Electronics, Electrical Engineering and Computer Science, Queen’s University Belfast, comments:
“There’s a well-known meme among network engineers: ‘It’s always DNS.’
“DNS (Domain Name System) is like the internet’s phone book, translating human-friendly website names into the numerical IP addresses that computers use to find and contact each other. When DNS fails, online services can become unreachable even though they’re still running- users, whether human or other machines, simply can’t find them.
“Reports indicate that a DNS resolution issue was the root cause of today’s AWS outage affecting data centers in Virginia, with AWS attributing this to an infrastructure fault rather than malicious activity. Enterprises rely on cloud providers like AWS to avoid managing physical servers themselves, eliminating the need to manage power, cooling, staffing, and security- which makes cloud services both immensely popular and cost-effective.
“However, this concentration can create systemic risk. When many enterprises depend on the same cloud infrastructure provider – whether AWS, Google, or Microsoft- a single large-scale fault can cascade across thousands of organisations simultaneously. Today’s outage disrupted a wide variety of sectors globally. This highlights a fundamental tension in modern cloud computing: the efficiency gains of centralisation come at the potential cost of creating single points of failure that affect vast portions of the internet simultaneously, whether through attacks or faults.”
Professor Michael Madden, Established Professor of Computer Science & Head of Machine Learning Research Group, University of Galway, comments:
“A large internet outage today (21 Oct 2025) demonstrates how the internet services that we all use daily are highly dependent on a few large cloud service providers. The outage is affecting services as diverse as Amazon, Snapchat, Fortnite, and the UK’s tax authority,
“All of these services are cloud-based and rely on Amazon Web Services (AWS), which is one of the largest providers of cloud computing. AWS is generally extremely robust, and is designed to continue to provide services when even when sections of its cloud computing network go offline, either unexpectedly or for routine maintenance.
“In this case, the cloud computers themselves have not gone down, but Amazon reports that the problem is with a small but essential piece of technology called DNS resolution. DNS, which stands for Domain Name System, translates recognisable computer names such as www.universityofgalway.ie to numerical addresses such as 163.24.112, that allow other computers on the internet to find them.
“You can think of DNS as being like the Contacts on your phone: you don’t have to memorise a friend’s phone number to phone them or send them messages, because your Contacts stores the number associated with their phone. And if your friend changes their phone, you just update this in your Contacts and can continue to reach them.
“Because one of Amazon’s DNS servers is not working, many parts of the AWS cloud computing network are temporarily unreachable, causing this failure. Amazon have not said what cause the DNS service failure, it could be anything from a hardware failure to a cyberattack.”
Declared interests:
Dr Domhnall Carlin: None
Professor Michael Madden: None