In the last few weeks we’ve seen outages from AWS and Cloudflare make huge parts of the internet suddenly… vanish. Not because someone kicked a cable out of the wall, but because of how deeply interconnected modern digital infrastructure really is.
The internet is built like this - and that’s the problem
You’ve probably seen that well-known cartoon showing a giant, wobbly tower labelled “All modern digital infrastructure,” balanced on a tiny block labelled “A project some random person in Nebraska has been thanklessly maintaining since 2003.”
It’s funny because it’s true.
The internet is a vast, sprawling system made up of layers and dependencies - and many of those layers rely on a shockingly small number of companies, tools, and services. When one of those foundational pieces falters, the ripple effect hits everyone.
And that brings us to AWS and Cloudflare.
Recently, both have had outages that felt like the internet itself was collapsing. To understand why, we need to take a slightly closer-but still friendly-look at what these companies actually do.
What AWS actually is: the internet’s utility grid
Most people think of Amazon as the place you get next-day deliveries.
But AWS (Amazon Web Services) is the world’s biggest cloud computing provider, which means it rents out things companies used to keep in their own server rooms:
- Servers (the computers running websites and apps)
- Databases (where your information lives)
- Storage (photos, files, analytics data)
- Authentication services (how you log into apps)
- Queues, APIs, machine learning tools, content delivery systems, and about 200 other building blocks
AWS is essentially:
“Why build your own infrastructure when you can rent Amazon’s?”
Millions of businesses - from one-person startups to global enterprises - run their systems on AWS. Many companies that look independent are quietly running everything on Amazon’s machinery.
So when AWS has a misconfiguration, capacity issue, or regional outage…
…it’s like the electricity grid going out.
Not just one website fails. Everything connected to that part of the grid blinks out too.
What Cloudflare actually is: the internet’s traffic system
Cloudflare doesn’t host websites.
It manages, routes, and protects their traffic.
Technically speaking, Cloudflare provides:
- DNS (the phonebook of the internet - it tells your browser where to go)
- CDN services (caching content all over the world so sites load instantly)
- DDoS protection (absorbing massive waves of malicious traffic)
- Reverse proxies that sit between users and servers
- Network routing and optimisation to keep the web running fast
In plain English:
Cloudflare is the air-traffic control system of the internet.
When Cloudflare goes down - even briefly - websites may still be “up,” but requests can’t reach them, or get lost, or get blocked for safety reasons.
So the outage feels global.
Why one problem cascades into chaos
This is where your illustration becomes such a powerful metaphor.
Modern internet infrastructure isn’t evenly distributed. It’s heavily concentrated on a handful of platforms:
- AWS
- Google Cloud
- Microsoft Azure
- Cloudflare
- A few core open-source libraries and protocols
These pieces form the bottom layers of the giant digital Jenga tower we all rely on.
So when one of them shakes, every layer above it wobbles.
Some businesses lose only their login system.
Some lose half their site.
Some lose everything.
Some lose the tools they depend on to diagnose the problem.
That’s why outages can look chaotic from the outside - because behind the scenes, they are.
Why this seems to be happening more often
Two reasons:
1. We rely on the internet for far more than we used to.
Ten years ago, if a backend cloud region went offline, your local coffee shop or HR tool might not notice.
Today, everything runs online.
So we notice outages faster - and they affect daily life more directly.
2. Internet infrastructure has become incredibly complex.
AWS and Cloudflare operate on mind-boggling scales:
- Hundreds of data centres
- Millions of servers
- Billions of requests per second
- Constant security threats
- Global routing decisions happening in milliseconds
A tiny misconfiguration in one region can cascade unpredictably.
It’s not fragility - it’s complexity interacting with scale.
Should we be worried?
Not really. In fact, the big takeaway is this:
Outages expose the dependencies most people didn’t know existed - but they also make the infrastructure stronger.
Every incident leads to new safeguards, more redundancy, better failover systems, and more resilient architecture.
The internet isn’t falling apart.
It’s evolving - and sometimes evolution looks messy.