Load Balancer
Easy+12 XPWatch one traffic cop spread requests evenly across a row of servers so none gets swamped.
No visualization loaded.
Watch
—
Press Run to begin.
When you have many servers, something has to decide WHICH one answers each request. That's the load balancer — a traffic cop in front of your servers. Press Run and watch it hand out requests one by one, taking turns.
▸What does 'round-robin' mean?
Just 'take turns in order.' Request 1 → server 1, request 2 → server 2, request 3 → server 3, then back to server 1. It's the simplest fair way to spread work, like dealing cards around a table.
▸Why not send everything to the fastest server?
Because then that one server gets swamped while the others sit idle — and if it falls over, you're stuck. Spreading the work keeps every server busy but not overwhelmed, and if one dies the balancer just stops sending to it.
▸Isn't the load balancer itself a single point of failure?
Good catch — it would be, so in real systems you run more than one balancer too. But conceptually, start here: one balancer, many servers, work spread evenly. That's the core idea.
▸Why does this need 'stateless' servers?
Because the balancer might send your next request to a DIFFERENT server. If servers remembered things privately, that would break. Stateless servers (which keep nothing private) can be swapped freely — exactly what the balancer needs.