A FRIENDLY COURSE

Learn AWS,
the Azure dev's way

Twelve visual chapters that take you from az login to aws sts get-caller-identity and beyond. Built for engineers who already know Azure - every concept anchored against an Azure equivalent, side-by-side code in both clouds, and honest call-outs where the mapping breaks down.

NVIDIA green = section accents Azure blue = Azure side AWS orange = AWS side Terraform purple = HCL snippets

Why AWS, coming from Azure?

You already ship on Azure. So why learn AWS? Three reasons that matter in practice:

~25%
Azure cloud share
~32%
AWS cloud share
~200
Azure services
~240
AWS services
2010
Azure GA
2006
AWS GA (S3)
If you care about Azure gives you AWS gives you
Market reachStrong in enterprise + Microsoft shops. AAD/Office integration is best-in-class.Largest cloud by revenue. Most third-party tooling targets AWS first.
Mental modelHierarchy: tenant → mgmt group → subscription → resource group → resource.Flatter: organization → account → region → resource. Tags do the grouping work.
IdentityAAD-centric. Users, groups, service principals, managed identities, all one system.IAM split: users + roles + policies, plus separate AWS SSO / Identity Center for humans.
NetworkingVNets with NSGs (stateful), no separate stateless layer by default.VPCs with SGs (stateful) and NACLs (stateless). Two-layer model takes adjustment.
ServerlessFunctions + Logic Apps + APIM + Event Grid.Lambda + Step Functions + API Gateway + EventBridge. Same shapes, different glue.
Ecosystem fitEnterprise IT, .NET shops, Microsoft 365 + Power Platform tie-ins.Startups, ML/data, anything Kubernetes-shaped (EKS), most CNCF projects test on AWS first.

The mental-model shift, at a glance

AZURE hierarchy Microsoft Entra tenant Management Group Subscription Resource Group Resource AWS hierarchy AWS Organization Organizational Unit (OU) AWS Account (no equivalent - use tags) Resource
Five-tier Azure → four-tier AWS. The Resource Group concept simply has no clean equivalent - in AWS, tags carry that weight. This single difference shapes how you organize everything.
The honest take: AWS is not "better than" Azure. They overlap by ~90%. The hard part of learning AWS as an Azure dev is unlearning Azure-specific reflexes - the cleanly inherited Management Group policies, the implicit Resource Group as a scope, the single AAD-everywhere model. AWS's design pushes more into IAM policies and account boundaries. Once you internalize that, most of the rest is renames.

The learning path

Twelve bite-sized chapters. Each builds on the previous, but each also stands alone if you need a refresher later. Foundations (1-3) before services (4-9) before synthesis (10-12).

FOUNDATIONS 1. Hello, AWS accounts, CLI 2. IAM users, roles, policies 3. VPC networking CORE SERVICES 4. Compute EC2, Lambda, ECS 5. Storage S3, EBS, EFS 6. Databases RDS, DynamoDB 7. Messaging SQS, SNS, EventBridge OPS & ADVANCED 8. Security KMS, Secrets, ACM 9. Observability CloudWatch, X-Ray 10. EKS Kubernetes on AWS 11. Serverless Lambda end-to-end 12. IaC Terraform, Orgs, SCPs Foundations = unskippable scaffolding. Core services = the daily-driver chapters. Ops & advanced = production muscle.
Reading order: top-to-bottom, left-to-right. You can dip into any chapter after finishing the foundations row.

All 12 chapters

1

Hello, AWS

Accounts vs subscriptions, regions, AZs, CLI install, aws configure vs az login, profiles, MFA. The "where do I click first" chapter.

2

IAM & Identity

Users, groups, IAM roles vs Azure managed identities, identity- vs resource-based policies, trust policies, STS assume-role, permission boundaries. The densest chapter.

3

Networking (VPC)

VPC vs VNet, subnets, route tables, IGW, NAT GW, SGs (stateful), NACLs (stateless), VPC peering, Transit Gateway vs Azure vWAN.

4

Compute

EC2 instance families, AMIs vs Azure images, Lambda vs Functions, ECS/Fargate vs ACI, EKS overview.

5

Storage

S3 buckets vs Blob containers, storage classes, versioning, lifecycle rules, EBS vs Managed Disks, EFS vs Azure Files.

6

Databases

RDS engines vs Azure SQL/PG flex, Aurora basics, DynamoDB vs Cosmos DB (partition keys, capacity modes).

7

Messaging & Events

SQS vs Service Bus, SNS vs Event Grid topics, EventBridge, Kinesis vs Event Hubs. Closest 1:1 mapping of any chapter.

8

Security & Secrets

KMS vs Key Vault keys, Secrets Manager vs Key Vault secrets, ACM vs App Service certs, GuardDuty intro.

9

Observability

CloudWatch logs/metrics/alarms vs Log Analytics/Azure Monitor, CloudTrail vs Activity Log, X-Ray vs App Insights. KQL → CW Logs Insights.

10

Kubernetes on AWS

EKS control plane, node groups vs AKS node pools, IRSA vs workload identity, AWS LB Controller, EBS CSI, Karpenter.

11

Serverless patterns

Lambda + API GW + DynamoDB + EventBridge end-to-end vs Functions + APIM + Cosmos + Event Grid. Cold starts, layers.

12

IaC & Multi-account

Terraform AWS provider, modules, AWS Organizations vs Management Groups, SCPs vs Azure Policy, Control Tower landing zones.

Start with Chapter 1 →

How to read this doc set

Every chapter follows the same 7-block shape. Once you know the shape, you can skim or deep-read any chapter without losing your place.

BlockWhat you'll find
1. Sticky navJump between chapters anytime. Highlighted = current chapter.
2. HeroChapter title, number, one-line subtitle.
3. Lead + TOC"Why this matters for an Azure dev" paragraph, plus jump-to TOC.
4. AzureAWS cheat tableThe chapter's standout block. Concept-by-concept mapping. Honest gap call-outs in italic red where AWS lacks a clean equivalent.
5. Concept sections3-6 deep-dives per chapter. Each has an ELI5 box, a diagram (where useful), side-by-side Azure / AWS code, gotchas, and engagement blocks (see below).
6. "Try it" mini-lab1-2 copy-pasteable commands. Every lab labeled with rough cost ($0, ~$0.01/hr, etc).
7. Quick-check quiz4-5 multiple-choice questions with click-to-reveal answers. Active recall before you move on.
8. Recap + nextBullet recap, pitfalls box, link to the next chapter.

Engagement blocks sprinkled inside concept sections

BlockWhat it is
Fun factTrivia and surprising defaults that explain AWS quirks. Dashed purple border. e.g., why us-east-1 hogs all the global services.
Etymology box"What's in a name?" - origins of AWS service acronyms (S3, STS, ARN, IRSA, SCP). Helps the abbreviations stick.
Bug huntBroken code or config - find the bug, then click to reveal the solution. Amber border. Trains your eye for real "access denied" debugging.
Cost trap puzzleA realistic deployment scenario that hides a 4-5 figure surprise bill. Spot the cause, then click to reveal the math and the fix.
Real-world incidentActual AWS disaster (Code Spaces shutdown, 2017 S3 outage) with dollar-amount badges. Makes the gotchas visceral.
Try it labCopy-pasteable commands marked with $0 / paid cost badges. Most are read-only and free.
Project CompassThe recurring build-along. Every chapter adds one slice of a fictional gfn-reports service. By chapter 12 you've wired up almost the whole AWS stack.
About Project Compass: This is the doc's continuity thread. Across all 12 chapters you'll build gfn-reports, a fictional NVIDIA service that reads GeForce NOW session telemetry, aggregates it in DynamoDB, exposes a query API via Lambda + API Gateway, runs in a dedicated VPC, encrypts with KMS, and ends up Terraform-managed in a multi-account org. Each chapter's Compass slice picks up where the previous left off. You can skip Compass if you only want the concepts - the chapters stand on their own - but following Compass gives you a real working system at the end.

Before you start

You'll get the most out of this if you:

Cost discipline: Every "Try it" lab is annotated with rough cost. Most are $0 (read-only API calls or terraform plan without apply). When a lab spends real money it'll say so up front.