1.1.4. System Architecture

WHG v4 runs on the same platform as its predecessor, extended rather than replaced.

Component	Role
PostgreSQL / PostGIS	store of record: accounts, datasets, collections, the attestation model, geometry
Elasticsearch (at Pitt CRC, behind WHG’s gateway)	the ~47-million-record place index, search, phonetic (sounds-alike) matching
Map tiles	pre-generated vector tiles for gazetteer coverage and boundaries
Django web application	the website, Atlas, the Collaborative Workbench and the APIs

The decision to keep this stack, rather than move to a graph database, is recorded in the database assessment below, together with the measurements behind it.

• 1.1.4.1. Database Technology Assessment
  • 1.1.4.1.1. TL;DR
  • 1.1.4.1.2. Executive Summary
  • 1.1.4.1.3. Data Volume Projection
  • 1.1.4.1.4. Option 1: ArangoDB
  • 1.1.4.1.5. Option 2: PostgreSQL + Extensions
  • 1.1.4.1.6. Option 3: Neo4j
  • 1.1.4.1.7. Option 4: Vespa
  • 1.1.4.1.8. Detailed Comparison Matrix
  • 1.1.4.1.9. Recommendation
  • 1.1.4.1.10. Next Steps
  • 1.1.4.1.11. Conclusion
  • 1.1.4.1.12. The database choice is a critical architectural decision that will shape WHG v4’s development timeline, operational requirements, and long-term sustainability.
  • 1.1.4.1.13. Addendum: 2026 reassessment