Learn
Size it, then price it
The practical half of system design: how many servers, which database, how big a cache, and what it all costs on AWS. Each page has a calculator that shows its working.
Back-of-the-envelope estimation: the numbers every system design starts with
The numbers and formulas behind system design estimates: requests per second from daily users, storage, bandwidth, cache size and latency.
How many users can one server handle? A worked estimate
Turn daily active users into requests per second, then into servers: the formula, worked examples from 10K to 10M users, and a calculator.
AWS instance sizes explained: what micro, large and xlarge actually mean
What t4g.micro, m7g.large and c7g.xlarge mean, how much CPU and RAM each has, what each costs, and how many requests per second it handles.
Database sizing: which RDS instance, how many replicas, and when to shard
Size an RDS for PostgreSQL database from reads, writes and working set: when to scale up, add read replicas or shard. With a calculator.
How much memory does Redis need? Sizing an ElastiCache node
Estimate Redis memory from your keys and values, add per-key overhead and ElastiCache's 25% reserve, and pick the right node size.
Is a CDN worth it? When CloudFront costs more than your servers
CloudFront charges per request and per GB. See when a CDN saves money and when it costs more than your servers, with a calculator.
Now break one yourself.
The first challenge takes about two minutes. No signup.