Sharding

Sharding splits one logical database across several machines, with each holding a distinct portion of the data — customers A to M here, N to Z there.

Unlike replication, which copies everything everywhere, sharding divides. That is what makes it the answer when a dataset or its write volume exceeds what one machine can hold, and it is the last resort for exactly that reason.

Everything gets harder afterwards. Queries spanning shards must be assembled by the application, transactions across shards lose their guarantees, and choosing the wrong shard key produces one overloaded machine and several idle ones. Exhaust indexing, caching, replication and a larger machine first — most systems never genuinely need this.