What is NoSQL?

For decades, "database" meant tables and SQL. Then the giant web companies hit walls — billions of users, data spread across continents, shapes that wouldn't sit still — and a family of alternatives emerged: NoSQL ("not only SQL").

Why tables weren't always enough

Relational databases are magnificent, but they make trade-offs:

• Fixed structure — every row has the same columns; changing that on a billion-row table is major surgery
• Scaling up, not out — classically they prefer one bigger server over many small ones
• Object mismatch — your app thinks in nested objects; tables are flat

NoSQL databases relax one or more of these in exchange for flexibility or horizontal scale.

The four families

Family	Data shape	Famous example	Sweet spot
Document	JSON-like documents	MongoDB	app data with nested, evolving shapes
Key-value	key → value pairs	Redis	caching, sessions, blazing speed
Wide-column	sparse giant tables	Cassandra	write-heavy data at massive scale
Graph	nodes + relationships	Neo4j	networks: friends, routes, fraud rings

This trail focuses mostly on document databases — the most common first encounter — then tours the others.

What's next

The lingua franca of document databases — JSON — in ten friendly minutes.