Cloud architecture patterns, explained simply
Cloud architecture patterns are proven ways to combine cloud services for a common job, such as a web app, an API or a document assistant. Each CloudBrum pattern shows a diagram, the AWS, Azure and Google Cloud service for every part, and what to decide first.
Reviewed . Starting points, not design reviews. CloudBrum is independent of AWS, Microsoft and Google.
Patterns
All 20 patterns are ready.
Generative AI and machine learning
Generative AI application
Build an app that uses large language models for chat, summaries, writing or code, with safety checks.
RAG assistant over your documents
Answer questions from your own files, with sources, using search plus a language model.
Voice assistant or voice bot
Let people talk to your app: speech in, a model or bot in the middle, speech out.
Machine learning project
Prepare data, train and track models, and serve predictions, with GPUs when you need them.
Applications and APIs
Web application with a database
A typical web app: hosting, a database, file storage, login, a CDN and email, with monitoring.
REST API back end
A secured API in front of your business logic and data, with login, queues and monitoring.
Mobile app back end
Sign-in, push messages, files and data for an iOS or Android app, behind an API.
Static website
A fast, low-cost site served from a CDN, with a domain, a certificate and a form or two.
Event-driven or serverless back end
Small functions that react to events, with queues and topics between them.
Microservices on containers
Containers on a managed cluster with a registry, load balancing, secrets and logging.
Video platform
Upload, convert, store and stream video to viewers, with login and a CDN.
Data and analytics
Analytics or data platform
Bring data into a lake, clean it, query it in a warehouse and show it in dashboards.
Real-time data pipeline
Take in events as they happen, process them in motion, then store or chart the results.
IoT platform
Connect devices, collect their readings, store time series and act at the edge.
Platform, security and operations
Secure cloud foundation
The accounts, access, network, logging and cost controls to set up before any workload.
DevOps toolchain
Source control, build and deploy pipelines, registries, infrastructure as code and monitoring.
Backup and disaster recovery
Copies of your data and a plan to keep running somewhere else when something fails.
Migrating servers and databases
Assess what you have, move servers and databases, link the networks and govern the new setup.
Regulated industries
Healthcare application
Protect patient data with encryption, audit logs and access control while serving your apps.
Payments or fintech app
Secure transactions with encryption, DDoS protection, audit trails, compliance and recovery.
Cannot find your case here? Describe what you are building in the service suggester.
Frequently asked questions
What is a cloud architecture pattern?
A cloud architecture pattern is a reusable layout of cloud services that solves a common problem, such as hosting a web app or answering questions from documents. It shows the parts, how they connect and what to decide, so you do not start from a blank page.
Do the patterns cover AWS, Azure and Google Cloud?
Yes. Each pattern names the main service for every part on all three clouds, taken from CloudBrum's service suggester, so you can compare them side by side.
Are these designs ready to use in production?
No. They are starting points. Your data, users, rules and budget decide the real design, so confirm details in the providers' documentation and have an architect review a production design.
How are the service names kept current?
Service names come from CloudBrum's service suggester data, which is checked against the providers' own pages and shows its check date. Services that were closed or renamed are flagged on each pattern page.