Contact Us

HERITAGE ALLIANCE · DATA & KNOWLEDGE

Heritage Data Infrastructure Architecture

Explore the sections below

Beneficiary Data Node Requirements

Source reference requirements for a participating local node. Confirm capacity, security and operational needs with the responsible institution before deployment.

Hardware & operating system

  • Quad-core AMD or ARM processor.
  • At least 32 GB RAM; 64 GB recommended.
  • SSD storage of 1 TB or more.
  • Ubuntu 24.04.
  • Separate server or virtual machine preferred over a shared partition.

Operations & maintenance

  • Source availability target: 99%; follow stricter local requirements where applicable.
  • Nightly backup is recommended in the source design.
  • Assign system administration and network support.
  • Maintain patches, Docker components, deployment software and server configuration.

cHDP Architecture Overview

The source combines local data nodes with coordinating network services. Individual-level records remain subject to the source institution’s controls; approved queries and outputs require governance review.

Network Node Server

A grouped service environment for node registration, authentication, metrics and research tools. Callout leaders explain components; they do not introduce additional service dependencies.

Reference network node server with Docker and research services. Five explanatory leaders identify node registration, metrics, authentication and visualisation components.
Adapted reference architecture. FastAPI registers nodes; Prometheus collects metrics; Grafana presents metrics; Keycloak handles authentication; Superset supports exploration.
Enlarge diagramDownload SVGDownload PNGSource diagram ↗
Source reuse licence not supplied; retain component-specific notices.

Data Node

Local ETL maps source information to OMOP CDM. Docker-hosted OHDSI components and federated learning tools surround the database in the source design.

Reference local data node: ETL flows into an OMOP CDM database with grouped OHDSI tools, Docker, Ubuntu and Vantage6. Callout leaders explain components.
Adapted reference architecture. Ubuntu and Docker host the ETL, OMOP CDM, OHDSI tools and Vantage6 components described by the source.
Enlarge diagramDownload SVGDownload PNGSource diagram ↗
Source reuse licence not supplied; retain component-specific notices.

Global Infrastructure

Representative EMR systems and local nodes are grouped with NHIC network servers and the project server. No country labels occur in the source technical diagrams.

Reference global architecture: three representative data sources and local nodes are grouped with two NHIC network servers and a project server. Brace-shaped connectors have no arrowheads.
Adapted reference architecture. Source systems, local OMOP nodes, NHIC network servers and a project server are arranged as in the original. The legend identifies node health, registration keys and aggregate results.
Enlarge diagramDownload SVGDownload PNGSource diagram ↗
Source reuse licence not supplied; retain component-specific notices.

Key Components

Local Data Nodes

Institutions control their own data environment and permissions.

Containerised Services

Docker groups reproducible software components for deployment.

Central Coordination

The reference hub coordinates participating nodes and approved aggregate exchange.

Analytics & AI

Analytical tools support research questions and modelling where approved and validated.

How Data Flows

  1. Local collection

    Data originates in an authorised clinical, laboratory, surveillance or survey system.

  2. ETL and harmonisation

    Local transformations map data into an OMOP CDM structure.

  3. Federated queries

    Authorised analyses run under local controls; approved aggregate results may be shared.

  4. Analysis and interpretation

    Research teams interpret outputs, assess disclosure risk and review findings before release.

Platform Capabilities

Capabilities described by the source design include timely evidence, impact modelling, federated queries, privacy-preserving research and a web portal. Heritage deployment and validation of each capability must be confirmed separately.

Technology Stack

Docker, R, Python, Jupyter, OHDSI/ATLAS, Vantage6, OpenMRS, OMOP CDM, machine-learning tools, SQL, CSV ETL and firewall controls appear in the source architecture. Technology names do not imply an active integration or a shared software licence.

Explore More

Data Catalogue Governance & Protection Ethics & Standards

Architecture diagram