Aller au contenu principal

Hub & Spoke Model

Understanding the distributed architecture of Synap

Overview​

Synap uses a Hub & Spoke architecture where:

  • Data Pod = Central Hub (data owner)
  • External Intelligence Services = External Spokes (intelligence providers via marketplace)
  • Protocol = Standardized communication (Hub Protocol V1.0)

Architecture Diagram​

Components​

Data Pod (Hub)​

  • Role: Central data repository
  • Ownership: User owns all data
  • Control: Initiates all Hub requests
  • Storage: PostgreSQL + R2/MinIO
  • Open Source: Self-hostable

External Intelligence Services (Spokes)​

  • Role: External intelligence providers
  • Ownership: Third-party services (marketplace)
  • Control: Respond to Data Pod requests
  • Storage: No user data storage
  • Services: Expert agents, specialized AI capabilities

Communication Flow​

1. User Request​

User sends a request through the app → Data Pod

2. Local Analysis​

Data Pod's local agent analyzes the request

3. Service Decision​

If complex intelligence is needed, Data Pod calls external intelligence service

4. Token Generation​

Data Pod generates a temporary access token

5. Data Request​

Hub requests specific data with the token

6. Processing​

Hub processes data with expert agents

7. Insight Return​

Hub returns structured insights

8. Event Creation​

Data Pod transforms insights into events

Benefits​

Scalability​

  • Hub can scale independently
  • Multiple Data Pods can connect
  • No single point of failure

Security​

  • Data Pod controls all access
  • Temporary tokens only
  • Complete audit trail

Flexibility​

  • Users can self-host Data Pod
  • Hub can evolve independently
  • Protocol enables interoperability

Protocol Standardization​

The Hub Protocol V1.0 ensures:

  • ✅ Type-safe communication (Zod schemas)
  • ✅ Secure authentication (JWT tokens)
  • ✅ Complete traceability (audit logs)
  • ✅ Standardized insights (HubInsight schema)

Next: See Hub Protocol Flow for detailed flow documentation.