IT WASN’T BUSINESS as usual for the internet this morning.
A range of well-known websites were temporarily down earlier today because of a problem with a shared service provider called Fastly.
Major media, government, entertainment and shopping sites could not be accessed, showing an error message to users who tried to visit their pages.
The Irish Times was temporarily down, as well as other news outlets internationally like the BBC, the Guardian, CNN and the Financial Times.
The Irish government website – gov.ie – was working as normal, but the UK government’s site was affected, along with Sweden’s social security service.
Amazon, Reddit and Pinterest were also among the websites impacted.
So how did it happen?
The reason so many websites were affected at once was because of a problem with a service provider the sites use called Fastly, which provides “content delivery services”.
“What happens is that websites are effectively replicated – they’re distributed in many different places, we have many different copies on the same website essentially,” Professor at University College Cork’s School of Computer Science Barry O’Sullivan told The Journal.
“So, in order to make sure that you can get onto your favourite website very very quickly, you would connect to a particular version of it that would be local to you,” Professor O’Sullivan explained.
It’s a bit like getting an item from a distributer instead of from a manufacturer, he said.
“What happened this morning is that the software that basically manages the delivery of that content was impacted, and what people see then is that the service is down.”
The internet is a “very physical thing”, and and if the distribution of information across the internet is interrupted, it appears as if it’s “failing”.
Internet users who tried to access the affected websites while they were down were shown an error message.
Messages like “Error 503 Service Unavailable” and “connection failure” appeared on screens when trying to connect to an affected site.
Access to the websites has been largely restored.
Shortly after 12pm, Fastly said that it had identified a configuration which had caused the problem and disabled that configuration.
“Our global network is coming back online,” the company said.
Fastly / Twitter
Fastly / Twitter / Twitter
Professor O’Sullivan said that over the next few hours, even though the issue has been fixed, it’s likely that there may still be “some lags and delays as the whole system reaches equilibrium again”.
Is this like the attack on the HSE?
In short, no.
“It’s unlikely to be something sinister. It could be, these things can be caused by sinister means, but there’s no evidence to suggest that it is in any way sinister, so it probably is a good old fashioned computer bug, a glitch or human error,” Professor O’Sullivan said.
“People are probably very aware now of services being down because of the recent HSE and Department of Health cyberattacks. This is not of that nature at all.”
However, the websites being down, even temporarily, does have other major impacts.
There can be a “significant financial loss”.
“It’s a very expensive thing,” Professor O’Sullivan said.
“All of the sales that should have happened on the Amazons of the world and the transactions that should have taken place didn’t take place.”
Some of those sales that were missed during the outage will happen now that they’re live again, but others may not, which can be costly for the businesses involved.
While their websites were down, some newspapers and news outlets found creative ways of continuing to deliver the news.
The Guardian’s UK Technology Editor Alex Hern shared updates on the website’s status on Twitter.
Through a series of tweets, Hern updated social media users on the internet outages, which has now been converted to an article on the Guardian site.
In the US, technology news outlet The Verge used another approach and published an article on the outages by using Google Docs, an online word processor.
However, a feature on Google Docs that allows other users to edit a page meant that people viewing the doc could also add content to it.
Dan Hett / Twitter
Dan Hett / Twitter / Twitter
Today’s outage caused quite a stir, but it isn’t the first time that the internet has been affected like this.
In July 2020, Cloudflare – a competitor of Fastly’s that also provides content delivery services – experienced a configuration error that caused an outage for 27 minutes.
The company said it saw traffic drop by around 50% across its network.
Major cities affected by the Cloudflare outage last year included Los Angeles, Washington DC, London, Paris, Moscow and Sao Paulo.
The Cloudflare incident was “not caused by an attack or breach of any kind”.
To embed this post, copy the code below on your site
have your say