The Canonical Architecture Model

The Canonical Architecture Model is the single editable source of truth for your entire software system in Avylo.

Unlike traditional whiteboard or drawing tools where components are mere graphical boxes and text labels, Avylo treats every component as a semantically rich entity within a structured architecture graph.

Model Entities & Topology

The Canonical Architecture Model models your system with precise architectural types:

1. Compute & Services

• Web Applications: Frontend single-page apps, SSR applications, mobile clients.
• API Gateways & Ingress: Reverse proxies, load balancers, rate limiters.
• Backend Services: Monoliths, microservices, serverless functions, background workers.

2. State & Storage

• Primary Relational Databases: PostgreSQL, MySQL, CockroachDB.
• Document & Key-Value Stores: Redis, DynamoDB, MongoDB.
• Analytical & Vector Stores: ClickHouse, Snowflake, pgvector, Pinecone.
• Blob & Object Storage: S3, Cloud Storage, R2.

3. Messaging & Integration

• Message Queues & Brokers: RabbitMQ, SQS, Celery, BullMQ.
• Event Streaming: Apache Kafka, AWS Kinesis, Redpanda.
• Third-Party Providers: Stripe, Auth0, Twilio, OpenAI, Resend.

Authoritative Model vs. Derived Views

A core design principle of Avylo is the separation between the Canonical Model and Derived Views.

       Canonical Architecture Model (Authoritative Source)
                               ↓
  ┌────────────────────────────┼───────────────────────────┐
  ↓                            ↓                           ↓
Visual Canvas            C4 Diagram Views            Exported Specs

• You edit the Model: When an Architecture Decision is accepted, the model updates.
• Views re-render automatically: The visual canvas, domain-specific zoom views, sequence diagrams, and architecture exports update deterministically.
• No out-of-sync diagrams: You never have to manually redraw connections or update box labels across multiple static drawing files.