Overview
Rent compute, storage, and network from a provider instead of buying servers. No cloud account lives in this tab.
On this page7 sections
The idea
Cloud computing means you rent computers, storage, and network from a provider instead of buying servers and putting them in a room. You open a browser, pick a region, and an API creates a machine, a bucket, or a database in minutes.
The provider owns the buildings, the power, the cooling, and the hardware. You pay for what you use: hours of a virtual machine, gigabytes in a bucket, bytes leaving the network. When you delete the resource, the bill for that resource stops.
No cloud account lives in this tab. The Python editor runs small dictionaries that stand in for cloud SDKs. You learn the words and the shape of the APIs here. Opening AWS, GCP, or Azure is homework you do on a real free-tier project later in the track.
Why this exists
An e-commerce site that stores orders on one office laptop loses the data when the laptop dies, and cannot serve a dashboard while the machine is asleep. A cloud bucket survives hardware failure because the provider copies objects across buildings. A warehouse query can scan yesterday's orders without anyone plugging in a new disk.
Data engineering jobs assume the lake, the warehouse, and the scheduler live in a cloud account. You will still write Python and SQL. The difference is that storage, compute, and network are APIs you call, not racks you install.
Picture this
Think of a hardware store that rents tools. You pay for the drill while you use it, return it when the job is done, and the store keeps it working. Cloud providers rent compute, storage, and network the same way.
Pipelines need all three. A bucket without compute cannot transform files. Compute without a network cannot reach the warehouse.
Compute is the CPU and memory that run your code: a virtual machine that stays up all day, or a function that wakes for one file. Storage is where bytes live when the machine is off: object buckets for files, disks attached to VMs, sometimes a managed database. Network is how those pieces talk: a private neighborhood (a VPC), DNS names, and rules for what may connect.
- You create a project or account with a provider (AWS, GCP, or Azure).
- You pick a region (a geographic cluster of datacenters) so data stays close to users or to legal requirements.
- You create storage, then compute that reads it, then network rules so only the job can reach the database.
- You delete or pause what you do not need so the bill does not keep running.
Same three jobs. Different purchase order.
| You used to buy | You now rent | Why data engineers care |
|---|---|---|
| A server in a closet | A virtual machine or a function | Jobs run without a laptop staying awake |
| Hard disks and NAS boxes | Object storage and managed disks | Bronze files survive one machine dying |
| Cables and a firewall appliance | VPC, subnets, and private endpoints | The warehouse is not on the public internet |
A region is a geography, not a vibe. Mumbai and us-east-1 are different buildings, different latency, and sometimes different legal rules. You pick one for a pipeline and keep bronze, compute, and the warehouse in that region unless you have a reason to copy data.
A small example
Model the three pillars as a Python list. The Python editor cannot create a VM. It can make the vocabulary something you can print and remember.
PILLARS = ["compute", "storage", "network"]
print("A cloud account rents:")
for name in PILLARS:
print("-", name)
print("count", len(PILLARS))A slightly richer model stores a one-line job for each pillar. You will use nested dicts like this when later lessons map AWS, GCP, and Azure product names.
CLOUD = {
"compute": "run the job",
"storage": "keep the files",
"network": "connect the pieces",
}
print(list(CLOUD.keys()))
print(CLOUD["storage"])Common beginner questions
Is the cloud just someone else's computer?
That joke is half true. You are still using computers. The difference is the API, the bill, and the fact that the provider copies data across buildings so one disk failure is not the end of the lake.
Do I need to know all three providers?
Depth in one and reading fluency in the others is the usual path. The next lesson names AWS, GCP, and Azure side by side so a posting that says S3 is not a different profession from Cloud Storage.
Can I learn this without a credit card?
Yes for this tab. Later, the capstone is a free-tier checklist you run in a real project. Do not paste access keys into the Python editor.
The bill does not pause itself
A VM you forgot to stop keeps charging. Object storage looks cheap until you keep years of bronze in the hot class. Delete or lifecycle what you do not need.
Before you start
This track assumes you completed Core Python. If you have not, start there first. Exercises here use lists and dicts, not a live cloud SDK.
What comes next
The next lesson names the main service types a data engineer uses: compute, storage, databases, networking, and IAM, with AWS, GCP, and Azure product names in one table.
Practice
Run Sample to print the three pillars. Then complete Exercise: assign ["compute", "storage", "network"] to result and print the list.
Practicals · load into the editor
After you read the theory, run these in the pane on the right. They execute in this tab, no cluster.