Have you often envied the rich and crisp AWS diagrams you see online, but aren't sure how to recreate them with confidence? The tips I've collected may help you improve your deployment diagrams. Get ready to deliver cloud solutions architectures that are robust, scalable, and cost-effective!
Prerequisites
Before jumping into drawing your AWS architecture diagrams, make sure you have the following prerequisites in place:
Declare Your Quality Attributes and NFRs
Ensure you cover all quality attributes and metrics in priority order. You can use a ready-made template to make your work efficient. Some are available publicly:
- You can look at the Arc42 quality model here; several examples are available in the template as well.
- Well-Architected pillars, for instance Azure or AWS, come with several examples.
Understand Higher-Level Technology Strategy and Context
If the organization's architecture is heading in a well-defined direction, your solution should follow the same approach.
Before working on the details of your own diagram, collect all documents that capture relevant technology strategy for your case.
The contextual view from the C4 model helps you concentrate on your surroundings and minimize effort on details. It is strongly recommended to initiate the first steps of your architecture work with this piece. The main focus should be:
- Actors using your system
- External systems as integration points
Narrow Down Your Architecture Styles and Patterns
Don't forget that this must be aligned with the higher-level technology strategy.
A comprehensive list of architecture styles and patterns can be found at:
- Azure architecture guide
- Wikipedia
- A comprehensive infographic on software architecture styles from ByteBtyeGo
Research
Collecting Use Cases
The following pages and guides hold valuable blueprints for typical use cases.
Finding comprehensive AWS use-case blueprints can be challenging, as they are often scattered across different AWS resources. Collecting these links in one place makes it easier to find proven architectures and patterns. This helps you quickly apply AWS best practices to your own projects.
As the first step, go through this list of resources to gather relevant use cases for your architecture:
- AWS Prescriptive Guidance – make sure you scroll all the way down and check "Guides," "Patterns," and "Strategies."
- AWS Reference Architecture Diagrams
- AWS Case Studies
- AWS Solutions Library – make sure you scroll all the way down and check both "Find solutions by category" and "Browse all solutions."
- AWS Architecture Center
- AWS Well-Architected Lenses – make sure you check these special well-architected lenses in the "AWS Well-Architected Lenses" section.
- AWS Whitepapers & Guides
General Advice Before Starting to Draw
Pay attention to the style and iconography used in the diagrams you find. Notice how groupings are used to distinguish different parts of the architecture, how to display multi-AZ, multi-region setups, and how to represent security boundaries.
Tools for Drawing
After finishing your research, it's time to start drawing your architecture diagrams. Here are some useful resources to help you recreate the AWS style and iconography.
Iconography
Here's a website that provides a full list of AWS icons:
Diagramming Tools
- DrawIO – my favorite tool, able to draw a whole set of different AWS architectures, including all icons. Limited support for 3D AWS deployment views as well.
- AWS Composer – able to generate ready-to-deploy IaC, and integrate with IDEs for generative AI-powered suggestions.
- CloudCraft – specialized in 3D AWS architecture diagrams with support for cost estimation.
Refine Your Deployment View
After finishing your first draft, it's time to review your work. The tools below help you ensure that your architecture follows AWS best practices and matches your requirements. Gradually refine and polish your diagram based on the materials below.
AWS Well-Architected Pillars
Make sure you deep dive into each pillar and collect those gudelines which are relevant for your deployment. Use these as a checklist to verify that your architecture meets AWS best practices. Each pillar has its own dedicated whitepaper. I'm going to list them all in here for your convenience:
- Overview of the Pillars
- Operational Excelence
- Security
- Reliability
- Performance
- Cost Optimization
- Sustainability
Evaluating Technologies
Open source projects to find libraries, tools, and resources for your services help narrow down technology choices.
Serverless Patterns
A full set of serverless patterns with example code and one-click deployment. You can use it to polish a specific area of your architecture if you're considering a serverless architectural style.
Justify and Explain Your Deployment View
There are some ready-made materials to help make your architecture more understandable. The C4 model is an excellent framework for architecture analysis, so I use most of their provided resources to cross-check my diagrams.
- Diagram review checklist
- techtribes.je – sample high-level architecture diagram from Simon Brown, showcasing some of the best practices from the diagram review checklist.
- Arc42 template – deployment view – The arc42 template offers high-quality, polished guidelines. This section provides specific advice on how to represent deployments.
Provide an Explanation for Your Deployment View
Make sure you back up your technology choices represented in the deployment view with explanations. For instance, you can use the "Architecture Decisions" section from the Arc42 template.
Don't Forget About Cost and Capacity Estimates
Crafting a good deployment plan is an iterative process. Make sure you revisit your architecture diagram and refine it based on cost and capacity estimates. For your cost estimates, you can use the AWS Pricing Calculator.
To get realistic price calculations, you need basically two things:
- Technology choices, to know which services you are going to use.
- Capacity estimates, to be able to fill in the planned workload.
Making good capacity estimates is an art in itself. I recommend a book that provides guidance on how to do it properly: The Art of Capacity Planning
If it turns out that one specific piece of your architecture is too expensive, you don't necessarily have to proceed with it. Instead, explore alternative services or deployment models to reduce costs.
Power of GenAI
If you want to speed up your work, you can use Amazon Q with AWS-provided MCP services as a CLI tool. The generated diagrams are often not enough to meet quality standards, but I think the generated output can be used for exploring different options. Also, with the help of MCP, Amazon Q is able to search inside AWS documentation and help you clarify differences between specific AWS services and deployment models.
Top comments (0)