Part 2 — Load Balancer Architecture and Internals

maang.io System Design Series

What is this?

In Part 1 we treated the load balancer as a friendly host at the door. Now let's go behind the door and see how it actually works — how it picks a server, how it knows which servers are alive, how it handles encryption, and how it avoids being the one fragile piece that takes everything down.

The one-line idea: a load balancer is a little pipeline — accept → read → choose a healthy server → forward — wrapped in machinery that constantly checks who's healthy and keeps a backup of itself ready to take over.