Overview
Compute, storage, databases, networking, and IAM. Same jobs on AWS, GCP, and Azure, with different product names.
On this page7 sections
The idea
A cloud account is a catalog of services, not a single product. Data engineers live in five of those aisles: compute (run code), storage (keep files), databases (SQL you do not install yourself), networking (private paths between resources), and IAM (who is allowed to do what).
AWS, GCP, and Azure sell the same five jobs with different product names. EC2, Compute Engine, and Virtual Machines are rented computers. S3, Cloud Storage, and Blob Storage are object files. You learn the job first, then the label on the box.
No console lives in this tab. A nested Python dict is the map. You look up a category, then a vendor, and you get the name you will see in job postings and docs.
Why this exists
A pipeline that lands orders needs all five. Compute runs the extract. Storage holds the bronze file. A database or warehouse answers SELECT for the dashboard. Networking keeps the warehouse off the public internet. IAM makes sure the job can read bronze and cannot delete production.
Postings mix vendor names freely. If you only memorized S3, a Cloud Storage interview feels like a different career. If you know object storage as a category, you translate in a sentence.
Picture this
Picture a hardware store with five aisles. Each aisle is a job (compute, storage, databases, networking, IAM). Each brand (AWS, GCP, Azure) puts its own label on the shelf. You shop by aisle, not by logo.
The rest of the catalog (CDN, email, IoT) exists. Night jobs spend most of their time in these five.
Compute is VMs, containers, and short functions. Storage is object buckets for files and disks attached to VMs. Databases are managed Postgres or MySQL, plus the analytics warehouses (BigQuery, Redshift, Synapse) you will meet later. Networking is VPC or VNet, subnets, and private endpoints. IAM is users, roles, and service accounts.
Same five jobs. Three nameplates. Learn the row, then the cells.
| Category | AWS | GCP | Azure |
|---|---|---|---|
| Compute | EC2 | Compute Engine | Virtual Machines |
| Object storage | S3 | Cloud Storage | Blob Storage |
| SQL database | RDS | Cloud SQL | Azure SQL |
| Network | VPC | VPC | VNet |
| Identity | IAM | IAM | Entra ID + RBAC |
You do not need every SKU. You need to recognize which aisle a new name belongs in. Lambda is compute (a function). Cloud Composer is compute plus scheduling. Secret Manager sits next to IAM: it stores the password IAM decides who may read.
A small example
Store the map as nested dicts. The walkthrough later in the track uses GCP. The compute row still lists all three vendors so an EC2 shop is not a surprise.
SERVICES = {
"compute": {"aws": "EC2", "gcp": "Compute Engine", "azure": "Virtual Machines"},
"storage": {"aws": "S3", "gcp": "Cloud Storage", "azure": "Blob Storage"},
"database": {"aws": "RDS", "gcp": "Cloud SQL", "azure": "Azure SQL"},
"network": {"aws": "VPC", "gcp": "VPC", "azure": "VNet"},
"iam": {"aws": "IAM", "gcp": "IAM", "azure": "Entra ID + RBAC"},
}
print(SERVICES["compute"]["aws"])
print([SERVICES["storage"][c] for c in ("aws", "gcp", "azure")])Walk every category and print the three names. This is the same loop you will use when a posting lists products you have not clicked yet.
for category, vendors in SERVICES.items():
names = ", ".join(f"{cloud}: {product}" for cloud, product in vendors.items())
print(f"{category}: {names}")Common beginner questions
Is a warehouse the same as a database?
A managed SQL database (RDS, Cloud SQL, Azure SQL) is for applications: orders, users, payments. A warehouse (BigQuery, Redshift, Synapse) is for analytics: large scans, dashboards, night jobs. Both are 'databases' in casual speech. They are different products and different bills.
Why does GCP also call the network VPC?
Vendors reuse words. AWS VPC and GCP VPC are the same idea (a private neighborhood). Azure says VNet. The next networking lesson in this track is enough to debug 'the job cannot reach the database.' It is not a networking certification.
Do I create all five on day one?
No. A first free-tier project is usually storage plus a warehouse. IAM comes with the project whether you notice it or not. Compute and a private network arrive when the pipeline leaves your laptop.
Do not memorize SKUs instead of jobs
If you only know 'S3' and a posting says 'Blob Storage,' you will freeze. If you know object storage, you ask how prefixes and IAM work on that vendor.
Python dicts, not a live SDK
boto3, google-cloud-storage, and azure-storage-blob are real libraries you will import on a laptop. Here, SERVICES is a dict so the Python editor can grade a lookup.
What comes next
The next lesson separates a data lake (cheap files) from a data warehouse (SQL tables), then names the lakehouse that uses both. Bronze usually lands in the lake.
Practice
Run Sample to print compute names from the SERVICES map. Then complete Exercise: collect EC2, Compute Engine, and Virtual Machines in that order and store the list in result.
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.
Practice this
Same ideas as interview drills. These challenges open in the studio with a dataset and tests already set up.