Category · Architecture

How to design a system that survives a million users

Step by step, the way a system design interview goes. Course plan: 1. The interview and requirements. 2. Scaling and the load balancer. 3. Caching. 4. Databases under load. 5. Queues and async processing. 6. CDN and static content. 7. API design and rate limiting. 8. Monolith or microservices. 9. Reliability and observability. 10. Case study: a URL shortener. 11. Case study: a news feed. The first ten lessons are available now, the news feed case study is coming soon.

10 lessons114 min

To the first lesson Cheatsheet now

Topics, each as its own lesson

01How a system design interview goes and where to startDemo

Requirements and numbers first, then the diagram. Not the other way round.

02Scaling: vertical, horizontal and the load balancer

A bigger server or more servers, and who spreads the requests between them.

03Caching: cache-aside, TTL and invalidation

How to serve hot data from memory without showing stale data.

04Databases under load: replicas, sharding, consistency

How to offload reads, split writes, and what you will have to give up.

05Queues and async processing

Not everything has to happen while the user waits for a response.

06CDN and static content

Images, scripts and video should come from a server close to the user, not from yours.

07API design and rate limiting

Clear resources, pagination, safe retries and protection from overly busy clients.

08Monolith, microservices or serverless

How to split a system into parts, and why microservices are not the starting point.

09Reliability and observability

SLOs, retries with backoff, circuit breakers and how to find out what exactly broke.

10Case study: a URL shortener

A classic interview problem from requirements to scaling, step by step.

0/10 topics open. This is the course intro, not the whole course on one page.

In the knowledge base

Reference pages for the technologies on this topic: who maintains them, when to use them and what interviews ask.