Why it matters

Sharding is what enables databases to scale beyond single machines. Understanding the trade-offs — and when to shard vs alternatives — is major architecture decision.

Advertisement

The architecture

Sharding key: column determining which shard a row belongs to. Common: user_id, tenant_id, geographic region.

Sharding strategy: hash (even distribution), range (localized queries), geographic.

Sharding designSharding keyuser_id / tenant_idStrategyhash / range / geoRebalanceon scale changesCross-shard queries are hard; design queries around sharding key when possible
Sharding parameters.
Advertisement

How it works end to end

Cross-shard queries: require distributed execution. Slow. Best avoided by design.

Rebalancing: adding nodes requires moving data. Consistent hashing minimizes movement.

Distributed transactions across shards need 2PC or sagas. Avoid when possible.

Alternative: managed distributed DBs (Spanner, CockroachDB, Yugabyte) handle sharding automatically.