Evaluate architectures
Architect scores ten reference architectures against a workload and your priorities, then presents a clear winner with evidence, alternatives, and tradeoffs. When a domain profile exists, the priorities start pre-tuned to your business context.
A saved domain profile pre-tunes the priority sliders. You can always override them — the sliders are suggestions, not lock-ins.
When to use it
- Choosing the foundation for a new workload or platform.
- Comparing messaging or integration patterns for a specific scenario.
- Re-evaluating a choice when priorities shift (try the What-If tuning).
The five steps
Select a workload scenario
Pick one of the 18 scenarios, e.g. E-Commerce Order Platform or Event Streaming Platform. Each scenario carries realistic requirements and quality guidance.
Review the priority sliders
Eight criteria — scalability, reliability, operational simplicity, cost efficiency, latency, ordering, replay, and compliance — start weighted from your domain profile. Adjust them to reflect the decision at hand.
Generate the recommendation
Architect evaluates all ten architectures and ranks them. The winner gets a score, a confidence level, and a plain-English rationale.
Explore the tradeoffs
Compare alternatives side by side, review the tradeoff matrix, and inspect the recommended architecture in the blueprint diagram.
Export or hand off
Download the blueprint and ADR for your records, or jump straight to Reviewer to score the design you actually plan to run.
The ten reference architectures
| Architecture | Category |
|---|---|
| Apache Kafka | Log-based streaming |
| Amazon Kinesis Data Streams | Log-based streaming |
| Google Cloud Pub/Sub | Managed pub/sub |
| Amazon SQS | Queueing |
| RabbitMQ | Queueing / broker |
| Amazon EventBridge | Event routing |
| Amazon API Gateway | API / event ingress |
| AWS Lambda | Serverless compute |
| AWS Elastic Container Service | Container compute |
| Red Hat OpenShift on AWS (ROSA) | Managed Kubernetes |
The eight criteria
- Scalability — headroom for growth and peaks.
- Reliability — durability and uptime guarantees.
- Operational Simplicity — how easy it is to run.
- Cost Efficiency — operating cost behavior.
- Latency — end-to-end delay.
- Ordering — message ordering guarantees.
- Replay — re-processing past events.
- Compliance & Governance — audit and regulatory fit.
Workflow tips
-
Deep links. A scenario can be opened directly with
?scenario=<id>, and the recommendation links into Reviewer with the chosen architecture pre-selected. - What-If tuning. Move a slider after generating a result to see how the ranking shifts — great for "what if cost matters more?" conversations.
- Read the Q&A. The built-in Q&A explains the tradeoffs behind each scenario in plain language.
Architect hands off to Reviewer automatically: the "Review this design" link opens Reviewer with the scenario and winning architecture pre-loaded.