When I told my dad I had started learning Azure, he asked me one question.
“Where did you start?”
I mentioned tutorials, some documentation, and browsing the portal.
He said something simple:
“Start with compute. Everything else in the cloud builds on top of it.”
At the time, it sounded obvious. Of course applications need computers to run.
But when I opened the Compute section inside Azure, I started seeing what he actually meant.
The moment cloud stopped being abstract
In tutorials, cloud computing often feels very conceptual.
You hear words like scalability, high availability, distributed systems, and managed services.
But those are still ideas.
When you open the Compute section in Azure, the abstraction disappears.
Suddenly you see things like:
- Virtual Machines
- Virtual Machine Scale Sets
- Images
- Availability Sets
At first glance, it looks like a lot of different services.
But underneath, they all revolve around one question:
Where does your code actually run?
Because before anything else in the cloud can exist - databases, APIs, storage, authentication - something has to execute the code.
And that something is compute.
Virtual Machines: the cloud version of “another computer”
The first thing that stood out to me was Virtual Machines.
I already knew the concept. A virtual machine is basically a computer running inside another computer.
But seeing it in the cloud made it feel different.
Instead of running a VM on my laptop, Azure was essentially offering me a computer somewhere in a data center.
CPU. Memory. Disk. Operating system. All configurable.
Which immediately made me realize something:
Running a VM isn’t just about deploying code. It means you’re responsible for the machine.
That includes things like:
- managing the operating system
- updating dependencies
- monitoring resource usage
- handling crashes
- planning for scaling
Suddenly cloud felt less like a magic deploy button and more like real infrastructure, And that's when it made sense for me...
Why abstraction exists in the first place
This also made me appreciate platforms like Vercel more.
When I deploy projects there, I never think about:
- CPU
- RAM
- operating systems
- load balancing
Those problems are handled for me.
Azure still provides those abstractions too, through services like App Service and Functions.
But it also lets you go deeper.
And that’s the interesting part.
Cloud platforms operate on layers of abstraction.
At the lowest layer, you have machines.
At higher layers, those machines disappear behind managed services.
Understanding compute is like seeing the foundation under the building.
The moment cloud felt more real
Before this, deployment always felt like the final step of building something.
You finish coding, push your repo, click deploy, and move on.
But seeing compute changed that perspective slightly.
Deployment isn’t the end. It’s the point where your code finally meets the machines that will run it.
And suddenly questions like these start to matter:
- How many machines are running my app?
- What happens if one crashes?
- What happens when traffic spikes?
- How much does it cost to keep these machines running?
I still have a lot to figure out
Ironically, I wasn’t even able to experiment much yet because I ran into an account issue while trying to explore the free tier.
So most of this realization came just from exploring the compute section and thinking through what these services represent.
But even that small exploration already made the cloud feel less mysterious and more tangible.
Where this journey goes next
This is the second step in a small series where I’m documenting what I learn while exploring Azure and cloud architecture.
Last time, I wrote about opening the Azure Portal and realizing how big the cloud ecosystem actually is.
This time, I realized something simpler:
Everything eventually comes back to machines.
Compute is where cloud stops being abstract and starts becoming engineering.
And now I’m curious:
If you’ve worked with cloud platforms longer than I have,
what was the moment when cloud finally “clicked” for you?
Was it compute, networking, containers, something else entirely?
Because I feel like I’ve only just found the first layer.
Top comments (0)