Lesson 03 · Slide 1
Lesson 03 — The Cloud, Demystified
AWS, GCP, Azure — what they actually sell, and why every team has a bill.
IT Landscape
Module 1
Lesson 03 · Slide 2
What "the cloud" is, in one sentence
Renting computers, storage, and services from someone else's data center — by the minute.
That's it. No mystical fog. Just somebody else's machines, billed by usage.
Lesson 03 · Slide 3
Who runs "the cloud"?
The big three hyperscalers:
- AWS (Amazon Web Services) — ~31% market share.
- Microsoft Azure — ~25%.
- Google Cloud (GCP) — ~12%.
Plus runners-up: Oracle Cloud, Alibaba, IBM, Cloudflare.
Note: share figures shift annually — see refresh-checklist.md.
Lesson 03 · Slide 4
What they sell, layer by layer
The stack
- SaaS — finished apps
Gmail, Slack, Figma
- PaaS — managed building blocks
databases, queues, runtimes
- IaaS — raw infrastructure
compute, storage, network
Different teams care about different layers.
An ops engineer lives in IaaS. A backend dev lives in PaaS. Everyone uses SaaS.
Lesson 03 · Slide 5
The flagship products
| Provider | Compute | Storage | Database | Functions |
|---|
| AWS | EC2 | S3 | RDS | Lambda |
| GCP | Compute Engine | Cloud Storage | Cloud SQL | Cloud Run |
| Azure | Virtual Machines | Blob Storage | Azure SQL | Functions |
Don't memorize. Recognize. You'll meet these names in every job listing.
Lesson 03 · Slide 6
What the console looks like
[AWS Console — Home]
☰ Services 🔍 Search 🔔 Region: us-east-1 ▾
Recently visited: EC2 · S3 · IAM · CloudWatch
[ Compute ] [ Storage ] [ Database ] [ Networking ]
[ Security ] [ Analytics ] [ AI/ML ] [ … ]
Hundreds of services in one menu. Most teams use 5–10.
Refresh: screenshot subject to refresh — see refresh-checklist.md L03.
Lesson 03 · Slide 7
The pricing model that surprises everyone
- By the second (or millisecond for Lambda).
- By the GB-stored — per month.
- By the GB-transferred out (egress!).
- By the API call.
You can leave one EC2 instance running and burn $200 in a month without noticing.
Lesson 03 · Slide 8
Free tiers exist
| Provider | Free for | Plus |
|---|
| AWS Free Tier | 12 months | Always-free tier |
| GCP Free Tier | 90 days + $300 credit | Always-free tier |
| Azure Free Account | 12 months + $200 credit | Always-free services |
Cap = enough to learn. Not enough to host a real product.
Lesson 03 · Slide 9
"Cloud-native" means…
Cloud-native
Built for cloud:
- Auto-scaling
- Ephemeral storage
- Managed services everywhere
- Pay only when running
Lifted in
An old app moved into cloud VMs:
- Still behaves like on-prem
- Gets the bill, not the elasticity
- "Lift-and-shift"
Most new companies are cloud-native. Most enterprises are mid-migration.
Lesson 03 · Slide 10
Where the cloud is going in 2026
- AI services baked in — AWS Bedrock, GCP Vertex AI, Azure AI Foundry. The cloud is now also an AI-model marketplace.
- More "serverless" / managed everything — fewer VMs, more "give us your code, we'll run it".
- Multi-cloud reality — most companies use 2+ providers, intentionally or accidentally.
Lesson 03 · Slide 11
Why this matters to non-IT readers
This is where the bill lives.
- Cloud cost is now a top-3 line item in many companies.
- "FinOps" (Cloud Financial Operations) is a real job title.
- Even non-engineers in the org get pulled into cloud-cost conversations.
Finance, procurement, product — all now read AWS invoices.
Lesson 03 · Slide 12
Recap
- Cloud = rented computers, billed by the second.
- Three big providers (AWS, Azure, GCP) plus niche ones.
- Pricing surprises matter — egress, idle resources, cross-region.
Coming up: Plan B / Module 2 — How software is actually built.