Skip to main content

Geo-Distribution & Multi-Region

TL;DR​

Geo-distribution: Deploy across multiple regions globally. Benefits: Low latency (serve from nearest region), disaster recovery, compliance (GDPR data residency). Challenges: Data synchronization, consistency, cost.

Core Concepts​

1. Why Geo-Distribution?​

BenefitImpact
Low latency150ms → 15ms (serve from nearby region)
High availabilityRegion outage → failover to other region
ComplianceGDPR requires EU data in EU
Disaster recoveryNatural disaster won't take down entire system

2. Architecture Patterns​

Active-Passive (Disaster Recovery)​

Use case: Disaster recovery (EU idle until failover)

Active-Active (Multi-Region Writes)​

Use case: Global app, write from nearest region

Challenge: Conflict resolution (two regions update same data)

3. Data Strategies​

Geo-Sharding​

US users → US database
EU users → EU database
Asia users → Asia database

Pros: Simple, compliant (GDPR)
Cons: Cross-region queries hard

Global Replication​

Write to US → replicate to EU, Asia (async)

Pros: Low-latency reads everywhere
Cons: Replication lag (eventual consistency)

4. DNS Routing​

Geo-based routing:

User in US → DNS returns US server IP
User in EU → DNS returns EU server IP

Latency-based routing: Route to lowest latency endpoint

Failover: If US region down, DNS routes to EU

Common Interview Questions​

Q1: "Why deploy in multiple regions?"​

Answer:

  1. Latency: Serve from nearest region (150ms → 15ms)
  2. Availability: Region outage won't take down system
  3. Compliance: GDPR requires EU data in EU
  4. Disaster recovery: Natural disaster resilience

Q2: "How do you handle data consistency across regions?"​

Answer:

  • Geo-sharding: Users tied to home region (no cross-region writes)
  • Eventual consistency: Async replication (accept lag)
  • Strong consistency: Sync replication (slow, rare)
  • Conflict resolution: Last-write-wins or CRDT

Q3: "How does Netflix serve globally?"​

Answer:

  1. Content: CDN + Open Connect (cache in ISPs)
  2. Metadata: Replicated across regions (eventually consistent)
  3. User data: Sharded by region (home region for writes)
  4. Result: Low latency globally, resilient to region failures

Quick Reference​

Benefits: Low latency, availability, compliance
Patterns: Active-passive (DR), active-active (writes)
Data: Geo-sharding or replication
Routing: DNS-based (latency, geo, failover)

Part 3 Complete! You now understand advanced patterns.

Next: Part 4: Interview Mastery - Framework and case studies to ace interviews.