Curriculum

Foundations

What actually happens between a client and a server, what a millisecond costs, and why you cannot have consistency, availability and partition tolerance at once. Start here if you can draw the boxes but not defend them.
Free
5 modules · 23h · 26 lessons · Beginner
How the Network Actually Works
5 hours
6 lessons
Beginner
How the Network Actually Works
Free
01.1
Latency Numbers & Estimation
4 hours
5 lessons
Beginner
Latency Numbers & Estimation
Free
01.2
The Machine Underneath
5 hours
5 lessons
Beginner
The Machine Underneath
Free
01.3
Consistency, Availability & Partitions
5 hours
5 lessons
Beginner
Consistency, Availability & Partitions
Free
01.4
API Design & Contracts
4 hours
5 lessons
Beginner
API Design & Contracts
Free
01.5
The drills in this track
01.1 · Packet Path

Route a request through DNS, TLS and three hops, paying the real latency cost of every choice you make.

01.2 · Envelope

Size six systems against a ticking clock. Order of magnitude is a pass; a decimal place is not required.

01.3 · Saturation

A service is at 100% of something. Find which resource, before the queue depth passes the point of no return.

01.4 · Split Brain

Partition a cluster and choose what to sacrifice. Every choice shows you the user-visible consequence.

01.5 · Contract

Ship four versions of an API to a client you cannot update. Break it once and the round ends.

Every track is free. Start anywhere.