CESMIIi3X COMPATIBLEOPEN SOURCE

Schema-Driven Knowledge Graph

45 SM Profiles on GitHub. Push JSON, the graph builds itself. No LLM, no manual configuration. Live data from MQTT, PostgreSQL, OPC-UA fused into one graph.

45 Profiles
1

WHAT exists

SM Profiles

45 type definitions with inheritance. Machine → CNC, IMM, Lathe. Attributes, relationships, KPIs.

65 Sources
2

WHERE from

Source Bindings

65 data source mappings. PostgreSQL tables, OPC-UA endpoints, MCP tools. Column → attribute mapping.

10 Syncs
3

HOW it flows

Sync Configs

MQTT, Polling, Kafka, Webhook, Manual. Real-time UNS subscriptions. Schema-driven, not code-driven.

OSF vs. CESMII SMIP

CapabilityCESMII SMIPOSF i3X
Type SystemSM Profiles (flat)SM Profiles + inheritance + KPI refs
Data ModelInstance modelKnowledge Graph (nodes + edges + embeddings)
Data SourcesPlatform-specific connectorsSchema-driven: PG, OPC-UA, MQTT, Kafka, MCP, REST
Live SyncPlatform pollingMQTT UNS + DB Polling + pg-notify + Kafka
Multi-Source FusionOne profile = one sourceOne profile, N sources (polymorphic edge resolution)
Impact AnalysisNot availableGraph traversal: cascade effects, critical paths, alternatives
KPI CalculationManual / externalSchema-defined, auto-calculated in KG Builder Phase 7
ConfigurationUI-basedJSON on GitHub — versioned, reviewable, CI/CD ready
APIREST on SMIPi3X REST on KG + Swagger UI + OpenAPI 3.0
VisualizationPlatform dashboards3D Force Graph + KPI overlay
🔗

Polymorphic Edge Resolution

targetIdProp: "machine_id" resolves to all 8 machine types automatically. Add a new type — all existing edges find it. Zero source schema changes.

🧬

Schema-Driven Inheritance

Machine parent defines 18 BDE attributes + 6 KPIs. All children inherit. InjectionMoldingMachine adds 90 process parameters. 112 redundant definitions eliminated.

📈

KPI as First-Class Citizens

KPIs defined in JSON schemas with Cypher formulas, thresholds, categories. Auto-calculated by the KG Builder. 140 KPI nodes across 20+ machines, live.

🌐

Source-Agnostic Graph

Data from SAP, CSV, OPC-UA, MQTT, Kafka — doesn't matter. The graph fuses everything. The i3X API queries the graph, never a source directly.

KG Build Pipeline

1
Type System
Indexes + inheritance
2
Instance Data
PG + OPC-UA + MCP nodes
3
Live Sync
MQTT + Polling + Kafka
4
Tombstone
Remove stale nodes
5
Embeddings
Vector search
6
Sensors
Auto-discover from MQTT
7
KPIs
Calculate from properties
Result
794K nodes, 1.4M edges

This site uses a cookie to remember your preferences. Analytics are anonymous and cookie-free. Privacy Policy