Scalability
What you do when one of everything stops being enough.
8 topics
Vertical vs Horizontal Scaling
Why the boring answer — buy a bigger machine — is right far longer than interviews imply, and what actually forces the switch.
core15 minStateless Services & Session State
Getting state out of the process so any instance can serve any request — and where that state goes instead.
core20 minScaling Reads
Replicas, caches and precomputation — the three ways to serve more reads, and the staleness each one introduces.
core20 minSharding & Partitioning
Splitting data across machines when one can't hold it — choosing the key, and living with the queries the key makes impossible.
core25 minHot Keys & Load Imbalance
Why an evenly sharded system still ends up with one machine at 100% — celebrity users, viral content, and the fixes for each.
core20 minRate Limiting & Backpressure
Protecting a system from more load than it can serve — the four algorithms, where the counter lives, and what the client should do about it.
core20 minMulti-Region Architecture
Running in more than one region for latency, availability or data residency — and why the database is the part that makes it hard.
deep25 minWhat Scale Actually Costs
Where the money goes in a large system, and why cost per request is a design constraint rather than a finance problem.
deep20 min