How HyperRoute Works

HyperRoute is built on a simple but powerful principle: composition happens at build time, not runtime. This architectural decision unlocks better performance, reliability, and security.

---

The Three Pillars

HyperRoute consists of three main components:

Core — The Runtime Engine

The high-performance GraphQL router that executes queries against your federated subgraphs.

• Designed for maximum throughput
• Ultra-low latency overhead
• Built-in security and rate limiting
• Hot reload without restarts

DevKit — The Build Tool

The CLI that manages the lifecycle of your schema deployments.

• Composes subgraph schemas into a supergraph
• Detects breaking changes before deployment
• Publishes immutable snapshots to your registry
• Pushes updates to routers atomically

Platform — The Control Plane

The optional dashboard for visibility and governance (enterprise feature).

• Real-time dashboards and metrics
• Approval workflows for schema changes
• Audit trail for compliance
• Role-based access control

---

Data Flow

Here's how a GraphQL query travels through HyperRoute:

What makes this fast?

• No runtime composition — The supergraph is pre-built by DevKit
• Query planning is cached — Repeated queries skip the planning phase
• Parallel subgraph calls — Independent data fetches happen simultaneously
• Connection pooling — Persistent connections to your subgraphs

---

Build Time vs Runtime

The key insight behind HyperRoute's architecture:

Build Time (DevKit handles)

• Schema validation
• Composition into supergraph
• Breaking change detection
• Immutable artifact creation
• Registry storage

Runtime (Core handles)

• Query validation
• Query planning & caching
• Subgraph execution
• Response merging
• Security enforcement

Why this matters:

• Security — Core never needs registry credentials
• Safety — Failed compositions never reach production
• Speed — Instant rollbacks (just point to a previous artifact)
• Simplicity — Router is stateless and horizontally scalable

---

Bring Your Own Storage

HyperRoute doesn't lock you into a proprietary registry. DevKit works with your existing cloud storage:

Configure once in devkit.yaml:

registry:
  type: s3
  bucket: my-schema-registry
  region: us-east-1

---

Deployment Modes

Local Development

Everything runs on your machine. Perfect for testing.

routerd                    # Start router
hyperroute push --local    # Deploy directly to local router

Production

DevKit manages the entire deployment: it publishes to your registry, updates the Platform, and pushes to your routers. The router performs an atomic hot swap—zero downtime, no restarts.

hyperroute push --env production --name users --schema users.graphql

Pro Tip: The same commands work in both modes. Only the --env flag changes.

---

Configuration Files

router.yaml

Controls how the router behaves at runtime:

server:
  port: 4000
  
security:
  max_depth: 15
  rate_limit: 100

telemetry:
  metrics: true
  tracing: false

devkit.yaml

Controls how DevKit builds and deploys:

registry:
  type: s3
  bucket: schemas
  
environments:
  production:
    routers:
      - https://api.example.com

---

Next Steps

• Security Guide — Learn about built-in protections
• Explore Platform — Dashboards, approvals, and governance