Load balancers & CDNs, explained at a chai stall

No diagrams, no jargon, just the analogy that finally makes two of the most important scaling concepts click.

Picture the busiest chai stall in the city. One person, one stove, a line out the door. That's a single server, and it's about to fall over. How do you fix it without the line collapsing into chaos?

Load balancer: more stoves, one queue

You add three more chai-wallahs and put one calm person at the front whose only job is to send each customer to whichever stove is free. That calm person is a load balancer. Customers don't know or care which stove makes their cutting chai — the line just keeps moving. The IP you connect to for a big website is usually that calm person, not the stove.

CDN: a stall on every corner

Now imagine the same chai is available at a stall on every street corner, so nobody walks far. That's a CDN (content delivery network) — copies of your static content cached at edge locations physically near each user. It's why a video or image loads instantly: it isn't crossing the planet from one origin, it's coming from the corner down your street.

Load balancer: spread the crowd across identical workers. CDN: move the content closer to the crowd.

Most of what people call “scaling” is some combination of these two moves — just wrapped in cloud pricing pages and intimidating diagrams.