Horizontal Scaling
Built for billion-scale operations with horizontal scaling.
Overview
Entitybase is designed from the ground up to scale to 1 billion+ entities and 1 trillion+ statements. The architecture separates storage (S3) from indexing (Vitess), enabling independent scaling.
Scaling Strategy
Storage Layer (S3)
- Immutable snapshots — Append-only writes avoid update hotspots
- Infinite capacity — Object storage scales automatically
- Cost-effective — Pay only for what you store
Indexing Layer (Vitess)
- Sharding — Scale horizontally by entity ID ranges
- Read replicas — Distribute read load
- Range-based ID allocation — Prevents write hotspots
Workers
- ID Generator — Range-based allocation for Q/P/L IDs
- Dump Worker — Parallelized entity exports
- RDF Streamer — Continuous RDF generation
Performance Targets
| Metric | Target |
|---|---|
| Entity Creation | 777K/day sustained |
| Read Latency | Sub-millisecond |
| Storage | 2.84B entities over 10 years |
Scaling Characteristics
- Write throughput — 10 edits/sec with 90% new entities
- Horizontal sharding — By entity ID ranges
- No single point of failure — Stateless API workers
See also: Architecture, Scaling Properties