System Design 101: Scalability Basics

How does Netflix serve millions of users at once? It's not magic; it's system design.

Vertical vs Horizontal Scaling

• Vertical (Scale Up): Buy a bigger server with more RAM/CPU. Easiest, but has a limit.
• Horizontal (Scale Out): Add more servers. Harder to manage, but infinitely scalable.

The Load Balancer

When you have multiple servers, you need a traffic cop. A Load Balancer (like NGINX or AWS ALB) distributes incoming user requests across your servers to ensure no single server is overwhelmed.

Caching

Reading from a database is slow. Reading from RAM is fast.

Redis is a popular in-memory cache. By storing frequently accessed data (like user profiles) in Redis, you can reduce database load by 90%.

Understanding these components is the first step to becoming a Senior Engineer.