Load Balancer Placement Patterns
What you'll learn: Where load balancers physically sit in your cloud architecture and why placement matters for different traffic flows.
Where Load Balancers Live
Just like traffic lights get placed at intersections (not in the middle of highways), load balancers are positioned at strategic points in your network architecture. Their placement determines what traffic they handle and how your application scales.
Three Core Placement Patterns
Internet-Facing Load Balancers sit at the edge of your virtual network, in public subnets with Internet Gateway access. They receive traffic directly from users on the internet and distribute it to backend servers in private subnets. This is like a hotel lobby receptionist who greets outside visitors and directs them to the right room.
Internal Load Balancers live entirely within your private subnets. They distribute traffic between application components that should never be directly accessible from the internet. Think of a restaurant's kitchen coordinator who routes orders between different cooking stations—customers never interact with this person.
Multi-Tier Patterns combine both types. Internet-facing load balancers handle user requests at the web tier, then pass traffic to internal load balancers that distribute work across application servers, which might connect to yet another internal load balancer fronting database replicas. Each tier uses security groups and private subnets to isolate traffic flows.
Why Placement Matters
Placing a load balancer in a public vs. private subnet affects routing (route tables), security (security groups and NACLs), and whether it receives an Elastic IP address. Internet-facing balancers need routes to Internet Gateways; internal ones route only within your VPC.
Key Takeaway: Load balancer placement—internet-facing, internal, or multi-tier—determines traffic flow boundaries and is fundamental to secure, scalable architecture design.