Designing on AWS

Design a production workload on AWS, choosing its compute, data, and integration services for their trade-offs rather than their marketing.

For: Developers and architects designing systems on AWS
Assumes: You've built and deployed web applications, and know the architecture basics that Developer to Architect covers in its first three stages. No AWS experience needed.

16 steps in 4 stages About 4 h 45 min of reading

Steps open in a new tab, so this page stays where you left it. Steps you've opened turn grey.

Stage 1 · Foundations

Thinking in cloud trade-offs

The trade-offs every AWS design is judged by, and the networking every AWS service assumes.

Guide 13 min
AWS Well-Architected Framework
The six pillars are the cloud's architecture characteristics. Every service choice later in the path trades among them.
Guide 23 min
Network Fundamentals for Developers
Subnets, routing, and load balancers all assume this. Cloud networking is ordinary networking with an API in front of it.

Stage 2 · Basics

The four things every AWS workload is built on.

Guide 21 min
AWS IAM: Identity and Access Management for Architects
Every request to AWS is an IAM decision, so identity comes before anything you build.
Guide 17 min
AWS VPC: Network Architecture
Where your workloads live and what can reach them. Most later services sit in, or next to, a VPC.
Guide 22 min
AWS Elastic Load Balancing for System Architects
The front door for most workloads, and the first place health checks, scaling, and TLS meet.
Guide 19 min
Choosing AWS Compute: EC2, Lambda, and Containers on ECS, EKS, and Fargate
Chooses between EC2, Lambda, and containers, and between ECS, EKS, and Fargate once containers win.
Guide 30 min
AWS Lambda for System Architects
The compute model that changes design the most: invocation models, retries, and cold starts shape everything around a function.
Guide 20 min
Amazon S3 for System Architects
The storage nearly every AWS system touches, and the cheapest home for data that doesn't need a database.

Stage 3 · Intermediate

Where each workload's data lives, and designing for the store you chose.

Guide 14 min
Choosing an AWS Database
Relational first, and when DynamoDB or a cache earns its place. The next two steps go deep on the likely answers.
Guide 20 min
Amazon RDS and Aurora for System Architects
What AWS manages for a relational database, what you still own, and when Aurora is worth it.
Guide 18 min
Amazon DynamoDB for System Architects
Designing around keys and access patterns instead of queries, which runs against relational habit.
Resource 1 min
Database Selection Matrix
Access patterns mapped to store types, for the workloads that need more than one database. Keep it open for your next data decision.

Stage 4 · Intermediate

Connecting services without coupling them: APIs, queues, and events.

Guide 15 min
AWS API Gateway for System Architects
The managed front for APIs: authorizers, throttling, and choosing the API type that fits.
Guide 18 min
Amazon SQS and SNS for System Architects
Queues and fan-out, the backbone of decoupled AWS systems, with the delivery and retry semantics you have to design for.
Guide 16 min
Amazon EventBridge for System Architects
Events routed by their content rather than by queue, for when producers shouldn't know their consumers.
Guide 15 min
AWS Step Functions for System Architects
Multi-step workflows with retries and error handling owned by the service, not written into your code.