Enterprise System Architecture

Distributed infrastructure serving 12M patients across 2,000 facilities with 99.99% uptime and sub-100ms query performance

Citus Distributed PostgreSQL
FHIR R4/R5 Integration
HIPAA + HITRUST Certified

Quick Reference: Architecture Components

Citus Distributed Database

PostgreSQL extension that shards data across worker nodes (1 coordinator + 3 regional workers) for horizontal scaling

HAPI FHIR Server

Open-source FHIR R4 implementation with 146 resource types and 56 database tables for healthcare data exchange

Monitoring Stack

Prometheus (metrics collection) + Grafana (visualization) + Redis (caching) for real-time system health tracking

Multi-Layer Architecture

Click each layer to explore the technical stack

Infrastructure Layer

Load Balancer

HAProxy with active-active HA pair

  • • 50,000 requests/second capacity
  • • SSL/TLS termination (Let's Encrypt)
  • • <1 second automatic failover

API Gateway

Kong with OAuth 2.0 / JWT authentication

  • • 10,000 req/min rate limiting
  • • Circuit breaker protection
  • • Real-time analytics

CDN / Edge Cache

CloudFront with 450+ global POPs

  • • 99.9% cache hit rate
  • • 10 Gbps sustained bandwidth
  • • DDoS protection (AWS Shield)

Application Layer

Next.js 15 + React 19 Frontend

Rendering

SSR, SSG, ISR with API routes

Performance

First Contentful Paint <0.8s, Lighthouse Score 98

Bundle Size

<200KB gzipped with automatic code splitting

HEDIS Engine

  • • 90+ HEDIS measures
  • • Real-time and batch processing
  • • <100ms per measure calculation
  • • 99.9% accuracy validated

MIPS Calculator

  • • Quality, Cost, IA, PI categories
  • • Real-time score calculation
  • • CMS certified algorithms
  • • Automated QPP submission

Care Gap Analyzer

  • • ML-powered gap detection
  • • Risk-based prioritization
  • • Automated intervention suggestions
  • • Real-time closure tracking

Platform Services

HAPI FHIR Server

HL7 FHIR R4/R5 compliant with 100M+ resources

  • • SMART on FHIR authorization
  • • Bulk data import/export ($import/$export)
  • • Subscriptions and webhooks
  • • Full audit logging (provenance tracking)
  • • CDS Hooks integration

n8n Workflow Engine

500+ automated workflows executing 10,000/hour

  • • Visual workflow builder
  • • Error handling and retry logic
  • • Webhook triggers
  • • Scheduled executions (cron)
  • • 200+ healthcare system integrations

Redis Cluster

128GB memory across 6 nodes, 1M ops/second

  • • Session management
  • • Real-time pub/sub messaging
  • • Application cache layer
  • • Rate limiting counters
  • • Leaderboards and rankings

Consul Service Mesh

  • • Service discovery and health checking
  • • Secure service-to-service communication
  • • Traffic management and routing

Prometheus

  • • Time-series metrics collection
  • • PromQL query language
  • • Alerting rules engine

Grafana

  • • Real-time dashboard visualization
  • • Custom metric queries and alerts
  • • Multi-datasource support

Data Layer

Citus Distributed PostgreSQL Cluster

Architecture: 1 coordinator node + 3 regional worker nodes with automatic sharding by tenant_id

# Citus Cluster Topology

Coordinator: hedis-postgres-prod (port 5432)

├─ Worker 1: US East Region

├─ Worker 2: US Central Region

└─ Worker 3: US West Region

Sharding Key: tenant_id (organization)

Replication Factor: 3x (high availability)

Performance Specifications

  • Capacity: 50TB total storage across workers
  • Read Performance: 100,000 QPS (queries per second)
  • Write Performance: 50,000 TPS (transactions per second)
  • Replication Lag: <100ms between nodes
  • Query Time: Sub-100ms for distributed queries

Key Features

  • • Distributed SQL queries across all workers
  • • Columnar storage for analytics workloads
  • • Real-time replication with automatic failover
  • • Point-in-time recovery (PITR) to any second

Citus Distributed PostgreSQL Architecture

Coordinator Node

Query Router - Port 5432

Worker 1 - US East

Active
5M
Patients
15TB
Storage
35K
QPS

Worker 2 - US Central

Active
4M
Patients
12TB
Storage
28K
QPS

Worker 3 - US West

Active
3M
Patients
10TB
Storage
22K
QPS
Total Cluster Capacity
12M
Patients
37TB
Storage
85K
QPS
99.995%
Uptime
Horizontal Scaling

Add workers to increase capacity linearly

Regional Distribution

Data locality reduces latency by 60%

Real-time Replication

<100ms lag between nodes

TimescaleDB

Time-series extension for PostgreSQL

  • • 7-year data retention policy
  • • 95% compression ratio on historical data
  • • 1 billion+ time-series data points
  • • 50+ continuous aggregates
  • • Automatic time-based partitioning

MongoDB Cluster

Document store for unstructured data

  • • 200+ collections
  • • 500M+ documents
  • • 3-node replica set for HA
  • • Sharding by document type
  • • Compound and text indexes

Apache Kafka

Event streaming platform

  • • 100+ topics for event types
  • • 1M messages/second throughput
  • • 30-day retention window
  • • 1000+ partitions for parallelism
  • • 3x replication factor

Defense-in-Depth Security Architecture

HIPAA, HITRUST CSF, and SOC 2 Type II certified infrastructure

Network Security

  • VPC with private subnets (no public internet access)
  • Network ACLs and Security Groups (least privilege)
  • AWS WAF for application-layer protection
  • DDoS protection (AWS Shield Standard)

Identity & Access

  • Multi-factor authentication (MFA) required
  • Role-based access control (RBAC)
  • Single Sign-On (SSO) with SAML 2.0
  • API key rotation every 90 days

Data Protection

  • Encryption at rest (AES-256)
  • Encryption in transit (TLS 1.3)
  • Database field-level encryption for PHI
  • Key management (AWS KMS with rotation)

Compliance & Auditing

  • HIPAA compliance with BAA
  • HITRUST CSF certified
  • SOC 2 Type II attestation
  • Full audit logging to SIEM (Splunk)

Incident Response

  • 24/7 SOC monitoring
  • Automated threat detection (SIEM rules)
  • 15-minute incident response SLA
  • Documented disaster recovery plan

Backup & Recovery

  • RTO (Recovery Time Objective): 1 hour
  • RPO (Recovery Point Objective): 5 minutes
  • Continuous database replication
  • 30-day hot storage, 7-year cold archival

System Performance Metrics

Real-world performance serving 12M patients across 2,000 facilities

Current System KPIs

MetricTargetCurrentStatus
System Availability99.99%99.995%✅ Exceeding
API Latency (p99)< 50ms45ms✅ Meeting
Page Load Time< 2s1.5s✅ Exceeding
Error Rate< 0.1%0.01%✅ Exceeding
Throughput10K req/s12K req/s✅ Exceeding
Data Processing1TB/day1.2TB/day✅ Exceeding

Scalability Headroom

ComponentCurrentMax CapacityUtilization
API Server Pods10 pods50 pods
20%
Database Connections5005,000
10%
Redis Memory12GB128GB
9%
Kafka Throughput100K/s1M/s
10%
Storage Capacity5TB50TB
10%

Infrastructure Headroom: System currently operates at ~10-20% capacity, providing 5-10x growth runway before scaling required

Ready to See This Architecture in Action?

Schedule a technical deep-dive with our Solutions Architects to explore how this infrastructure serves your specific needs

Schedule Architecture ReviewCompare to Legacy Systems