Why deploy CrateDB at the edge
- Ultra-low latency: process and query data locally for immediate insights.
- High throughput ingestion: handle millions of events per second from IoT sensors, machines, or gateways.
- Offline resilience: continue operating even with limited or intermittent connectivity; sync when back online.
- Lightweight deployment: install on minimal hardware or embedded systems, while keeping full SQL functionality.
- Unified data model: manage time series, geospatial, JSON, and relational data in one database.
- Edge-to-cloud continuity: seamlessly replicate or aggregate data into your on-prem or cloud clusters for centralized analysis.
Architecture overview
CrateDB’s shared-nothing, distributed architecture makes it ideal for edge environments.
- Each node can independently store and process data.
- Automatic sharding ensures even data distribution across limited hardware.
- Columnar storage enables efficient aggregations and queries, even on small devices.
- Local-first design minimizes dependency on cloud roundtrips.
Typical use cases
- Industrial IoT and manufacturing: Monitor machinery in real time for predictive maintenance and production optimization.
- Smart mobility and logistics: Analyze vehicle telemetry and routing data locally to optimize operations.
- Energy and utilities: Process sensor data for grid stability and local forecasting.
- Telecommunications: Monitor edge nodes and user activity data with low latency.
- Smart buildings and cities: Enable real-time event processing for security, energy, and environmental systems.
Key capabilities
- Real-time ingestion and SQL analytics on mixed data types.
- Time series and geospatial support for sensor and location data.
- Dynamic schema for evolving edge data structures.
- Automatic indexing and partitioning for fast local queries.
- Works with lightweight containers and Kubernetes at the edge.
- Integrates with AI and ML platforms for on-device intelligence.
Deployment models
CrateDB can run:
- On edge gateways or industrial PCs, ingesting data directly from PLCs or sensors.
- On embedded systems, supporting ARM architectures.
- In micro-clusters, colocated near production lines or connected devices.
