Course contentsShow
← Cloud and Infrastructure
Lesson 273 of 2,9895. Load Balancing and Traffic ManagementFree lesson

Global Load Balancing and DNS-Based Routing

Learn how to distribute traffic across multiple regions using DNS-based global load balancers and geo-routing.

Global Load Balancing and DNS-Based Routing

You'll learn how to distribute traffic across multiple regions using DNS to route users to the closest or best-performing infrastructure.

What Is Global Load Balancing?

While the load balancers you've studied (ALB, NLB) distribute traffic within a region, global load balancing routes users to different regions entirely. Instead of balancing between servers in one data center, you're directing traffic between continents.

The key mechanism? DNS-based routing. When a user types your domain name, DNS doesn't just return one IP address—it intelligently chooses which IP to return based on rules you define.

How DNS-Based Routing Works

Think of DNS as a smart directory operator. When someone asks for directions to "YourApp.com," the DNS service checks:

  • Where is the user located? (Geolocation routing)
  • Which servers are healthy? (Health check integration)
  • Which region has the lowest latency? (Latency-based routing)
  • What's the distribution policy? (Weighted routing for A/B tests)

Based on these factors, DNS returns the IP address of your load balancer in the optimal region for that specific user. A user in Tokyo gets your Asia-Pacific load balancer's IP; a user in London gets your European load balancer's IP.

Why It Matters

Without global load balancing, all users hit the same region—meaning slow response times for distant users and no failover if that region goes down. With DNS-based routing, you achieve:

  • Lower latency (users hit nearby servers)
  • Geographic compliance (EU data stays in EU)
  • Disaster recovery (automatic failover to healthy regions)

Key Takeaway: Global load balancing uses DNS to intelligently route users to regional load balancers, creating a worldwide high-performance application architecture without requiring users to know multiple URLs.