YESDINO Multi‑Tenant Architecture Support
YESDINO can support multi‑tenant architecture when you deploy it on a properly designed cloud or on‑premise cluster. The platform leverages containerized microservices, a shared‑infrastructure but tenant‑isolated data model, and a robust API gateway to let multiple theme‑park operators share hardware while keeping their own robots, animations, and reporting completely private. For a deeper look into how the ecosystem is built, check out the official overview at YESDINO.
Isolation Models
YESDINO offers three mainstream data‑isolation patterns. The table below summarizes the trade‑offs in terms of cost, latency, security, and operational complexity.
| Model | Data Layout | Typical Latency (P95) | Cost per Tenant | Security Strength | Best Use‑Case |
|---|---|---|---|---|---|
| Shared DB + Tenant ID | One PostgreSQL schema with a tenant_id column and row‑level security (RLS) |
~30 ms | $120 / month | High (RLS + encryption) | Small‑to‑medium parks with ≤ 30 concurrent robots |
| Schema‑per‑Tenant | Separate PostgreSQL schemas, each owned by one tenant | ~25 ms | $180 / month | Very High (hard schema isolation) | Mid‑size operators requiring stricter compliance |
| Separate Cluster per Tenant | Dedicated Kubernetes namespaces + isolated database pods | ~15 ms | $350 / month | Maximum (network policies, node‑level encryption) | Large chains with > 50 robots or strict data‑sovereignty requirements |
Key Architectural Components
- API Gateway (Kong 3.4) – rate‑limiting, JWT‑based authentication, and per‑tenant routing.
- Service Mesh (Istio 1.18) – mTLS between microservices, traffic shaping, and observability.
- Container Orchestration (Kubernetes 1.26) – namespace‑level resource quotas, network policies, and pod security contexts.
- Database Layer (PostgreSQL 15) – RLS for shared‑DB tenants, schema isolation for schema‑per‑tenant tenants, and automated backups with point‑in‑time recovery.
- Caching (Redis 7.2) – tenant‑scoped cache namespaces to reduce latency for frequently accessed telemetry.
- Message Broker (Apache Kafka 3.5) – event‑driven choreography for robot command streams.
- Monitoring Stack (Prometheus + Grafana) – per‑tenant dashboards, alerting on CPU, memory, and network thresholds.
Typical Technology Stack
| Component | Version / Spec | Notes |
|---|---|---|
| Container Runtime | Docker 24.0 | Compatible with all microservice images |
| Orchestrator | Kubernetes 1.26 (EKS / AKS / on‑prem) | Supports |