Deployment Modes

HyperMesh supports two deployment modes that can be used independently or together in a hybrid pattern.

Virtual Mode

“Reason in seconds. Data stays in place.”

The Virtual Hypergraph Engine compiles hypergraph patterns into optimized native SQL/API queries and pushes compute down to existing data warehouses (Snowflake, Databricks) or clinical systems (Epic, Cerner).

Aspect	Detail
Latency	Seconds (batch)
Data movement	Zero-copy — data never leaves the source
Governance	Inherits existing warehouse RBAC and audit logs
Best for	Retrospective screening, batch enrichment, compliance-sensitive workloads

Configuration

hypermesh.yaml

mode: virtual
sources:
  - type: snowflake
    account: "org-account"
    warehouse: "COMPUTE_WH"
    database: "SECURITY_EVENTS"
    schema: "RAW"
  - type: databricks
    host: "adb-123456.azuredatabricks.net"
    token_env: "DATABRICKS_TOKEN"
    catalog: "main"

Native Mode

“Act in milliseconds. Full compute surface.”

Data is ingested into HyperMesh’s binary indexes (TPI/FMI/PSI) for sub-millisecond temporal range queries and real-time analytics.

Aspect	Detail
Latency	Microseconds to low milliseconds
Data movement	Ingested into local binary indexes
Compute	Full hypergraph algorithms, streaming analytics
Best for	Real-time alerting, live clinical decision support, streaming threat narratives

Configuration

hypermesh.yaml

mode: native
storage:
  dir: /data/hypermesh
  bucket_seconds: 60
  ttl_days: 90
  autocompact:
    threshold: 10000
    interval_seconds: 300

Hybrid Pattern

Enterprise customers typically use both modes together:

1. Virtual for exploration and batch analysis — query existing data without migration
2. Native for real-time production — ingest high-velocity streams for sub-millisecond alerting

┌─────────────────────────────────────────────────┐
│                  HyperMesh Engine                │
├────────────────────────┬────────────────────────┤
│     Virtual Mode       │      Native Mode       │
│  (Snowflake, Epic)     │   (Binary Indexes)     │
│  Batch enrichment      │   Real-time alerting   │
│  Retrospective scan    │   Live dashboards      │
└────────────────────────┴────────────────────────┘

Docker Deployment

See the Docker guide for container images, Compose files, and Kubernetes manifests.

Cloud-Managed (Coming Soon)

HyperMesh Cloud will offer fully managed deployments with:

• Auto-scaling based on query load
• Multi-region replication
• Managed backups with point-in-time recovery
• SOC 2 and HIPAA compliance built-in

Contact sales@hypermeshai.com for early access.