Why does the internet keep crashing so often? First Google Cloud, then AWS, now Cloudflare.
On November 18, 2023, a significant outage at Cloudflare disrupted access to numerous major platforms, including OpenAI, Spotify, X (formerly Twitter), and Canva. This incident is part of a troubling trend, as it follows a series of outages affecting major cloud service providers like Microsoft Azure and Amazon Web Services (AWS) in recent months. While the internet may feel increasingly unstable, experts assert that the frequency of outages has not risen; rather, the consolidation of digital infrastructure means that when one provider experiences issues, the impact is felt across many services.
Cloudflare is a leading Content Delivery Network (CDN) that plays a crucial role in managing web traffic and protecting websites from cyberattacks. According to Angelique Medina, Head of Internet Intelligence at Cisco ThousandEyes, CDNs serve as the “front door” for websites, allowing users to access content efficiently. However, when this “front door” is compromised, users lose access to a multitude of sites and applications. This is particularly concerning given that many popular platforms rely heavily on a small number of service providers like Cloudflare, AWS, and Google Cloud Platform (GCP). Ramutė Varnelytė, CEO of IPXO, emphasized that the digital economy’s reliance on a few key infrastructure providers exposes it to significant vulnerabilities, as a single malfunction can lead to widespread disruption.
Despite the perception of increasing outages, Cisco ThousandEyes reports that the overall frequency of service disruptions has remained stable. The key issue lies in the growing dependency on a limited number of cloud service providers, which amplifies the effects of individual outages. As more websites and applications consolidate their operations under these major players, the shockwaves from outages become more pronounced, leading to a cascading effect on internet accessibility. This situation raises important questions about the resilience of the internet and the need for diversification in digital infrastructure to mitigate the risks associated with such outages.
https://www.youtube.com/watch?v=IXpvUD5SDzA
Cloudflare suffered a
major outage
on Nov. 18, and it took many major platforms down with it. OpenAI, Spotify, X, and Canva were among just some of the websites and services that went down Tuesday morning.
Feel like you’re experiencing deja vu? Well, you’re not.
The recent Cloudflare outage is not to be confused with the recent
Microsoft Azure outage
, which resulted in Xbox, Minecraft, and other platforms going down. And don’t forget the other major incident
that occurred in October
, when issues at Amazon Web Services (AWS) led to Amazon, Reddit, Snapchat, and other big platforms going offline. It’s also completely separate from the June outage, when Google Cloud Platform (GCP)
and
Cloudflare took down swaths of the internet
OK, so the internet’s biggest websites and platforms have been going down a lot lately. What’s going on here?
Why do so many websites go down at the same time?
Those big social media platforms, e-commerce websites, and gaming networks are all using the same few major service providers to host and deliver their content, with Cloudflare being one of them.
Cloudflare is one of the largest Content Delivery Networks (CDN), which is a system that manages web traffic. When Cloudflare functions normally, it manages web traffic and protects platforms from attacks.
“CDNs are distributed infrastructures that accelerate content delivery and enhance user experience by caching and serving web content closer to users,” explained
Angelique Medina
, Head of Internet Intelligence at the network intelligence company Cisco ThousandEyes, to Mashable. “They essentially serve as the ‘front door’ to websites and applications, with users connecting to Cloudflare’s servers instead of those of its customers.”
“When this ‘front door’ becomes unavailable, users lose the entry point to many sites and applications,” Medina said. “Given the large number of customers and sites that they service, any meaningful disruption in their ability to deliver content could effectively render large parts of the internet unreachable to users.”
Basically, the internet has consolidated, and we’re all reaping the rewards of that consolidation.
“This isn’t just another technical setback,” said Ramutė Varnelytė, CEO of
IPXO
, a leading IP resource management platform in Europe, in a statement provided to Mashable. “It’s further proof of how exposed [the] digital economy is to various malfunctions in just a few of the digital infrastructure service providers.”
Are internet outages occurring more frequently?
Back in the earlier days of the internet, there were countless web hosting providers, and many companies even ran their own dedicated servers. Now, companies big and small are all utilizing the same few cloud service providers, whether it be AWS, GCP, or Azure. Cloudflare makes the internet even more vulnerable to disruption, as the company has no one-to-one direct competitor.
“The incident further emphasizes the inherent danger of the entire internet infrastructure relying on a few service providers,” Varnelytė said.
But, is that it? And are these services really going down more than usual? According to Cisco ThousandEyes, which
tracks outages
and
maps them
, that’s correct. Web downtime is actually not happening more than usual, even if it
seems
that way.
“Cisco ThousandEyes has not seen an increase in the frequency of service outages in cloud and other Internet infrastructure providers; however, the number of sites and applications dependent on these services has increased,” Medina shared. “Because these services are increasingly a point of centralization — with a small number of companies handling large swaths of the Internet — if something goes wrong, it can lead to the disruption of many sites and applications around the globe.”
In short, the number of disruptions isn’t increasing. Rather, the shockwaves of individual incidents are getting bigger.
