Federated Data System
Federated data discovery for healthcare, life sciences, and research consortia. Cross-institutional search with data in place.
Problem
Research and healthcare consortia need to search across institutions’ datasets without centralising or copying sensitive data. Identity, trust, and access must work across boundaries while respecting HIPAA, GDPR, and institutional policies.
Solution
A production-grade federated platform: React portal (search, compute, topology, admin), IdP and Trust Anchor (JWT/JWKS, RS256), federation API (metadata aggregation, search, DRS, access requests), and distributed data nodes. Semantic harmonisation and geofence engine support compliance. Docker Compose and Helm for deploy.
Why it matters
Federated discovery keeps data sovereign while enabling cross-border and cross-institution discovery and compute. Critical for clinical trials, genomics, and multi-site research.
Tech choices
- Portal — React SPA (Vite) for landing, search, admin.
- Identity — IdP (login, token exchange), Trust Anchor (JWKS, signing/verification).
- Federation API — Python service: metadata harvest, search, policies, DRS, PostgreSQL.
- Data nodes — Per-institution services (direct download, pre-signed URL, constrained API).
- Deploy — Docker Compose, Helm charts for Kubernetes.