Multi-Region Deployment on AWS

Multi-Region Deployment on AWS

Multi-region deployment provides the highest level of availability and lowest latency for global applications. This guide covers active-passive and active-active patterns, data replication strategies, and the operational complexity you need to be prepared for.

Active-Passive (DR-Focused)

One region handles all traffic; the second region is ready for failover:

Primary Region (us-east-1):
  Route 53 → ALB → ECS/EKS → Aurora Primary
  (Serves all traffic)

Secondary Region (eu-west-1):
  Route 53 → ALB → ECS/EKS (scaled down) → Aurora Read Replica
  (Standby, receives replicated data)

Failover:
  Route 53 health check fails → DNS switches to eu-west-1
  Aurora Global failover → Replica becomes primary

Active-Active (Performance-Focused)

Both regions serve traffic simultaneously, routing users to the nearest region:

Route 53 Latency-Based Routing:
  US users → us-east-1 → ALB → ECS → Aurora (local writes)
  EU users → eu-west-1 → ALB → ECS → Aurora (local writes)

Data Replication:
  Aurora Global Database: ~1 second replication lag
  DynamoDB Global Tables: sub-second replication
  S3 Cross-Region Replication: minutes

Data Replication Options

Conflict Resolution

Active-active with writes in both regions introduces data conflicts:

• DynamoDB Global Tables — Last writer wins (automatic conflict resolution)
• Aurora Write Forwarding — Writes in secondary are forwarded to primary (avoids conflicts)
• Application-level — Vector clocks, CRDTs, or custom merge logic

DNS and Traffic Routing

# Latency-based routing
aws route53 create-resource-record-set \
  --hosted-zone-id Z123 \
  --name app.example.com \
  --type A \
  --set-identifier us-east-1 \
  --region us-east-1 \
  --alias-target DNSName=us-alb.xxx.elb.amazonaws.com

aws route53 create-resource-record-set \
  --hosted-zone-id Z123 \
  --name app.example.com \
  --type A \
  --set-identifier eu-west-1 \
  --region eu-west-1 \
  --alias-target DNSName=eu-alb.xxx.elb.amazonaws.com

# Health checks for automatic failover
aws route53 create-health-check \
  --caller-reference us-east-1-health \
  --health-check-config Type=HTTPS,FullyQualifiedDomainName=us-alb.xxx.elb.amazonaws.com,Port=443,ResourcePath=/health

Operational Considerations

• Infrastructure as Code — Terraform/CloudFormation must deploy identically to both regions
• CI/CD — Deploy to both regions simultaneously or with canary rollout
• Configuration management — Region-specific settings (endpoints, keys) managed separately
• Monitoring — Centralized monitoring that covers both regions
• Testing — Regular failover testing to validate recovery procedures

Cost Impact

Multi-region deployments roughly double infrastructure costs:

• Compute: 2x (full capacity in both regions for active-active)
• Database: 1.5-2x (Global Database adds reader cost + replication)
• Data transfer: Cross-region transfer at $0.02/GB
• Optimization: Use smaller instances in secondary for active-passive to reduce costs

Eazy SaaS Tip: We recommend active-passive for most SMBs (80% cost savings vs active-active with minutes-level failover). We reserve active-active for applications with strict latency requirements for global users or regulatory requirements for data locality.