Sharding — Split Your Database

11 · Data Partitioning

Sharding — Split Your Database

One database can't hold the world's data. Sharding splits data across multiple database nodes. Each node holds a slice — together they scale infinitely.

ALL WRITE REQUESTS

SHARD KEY HASH FUNCTION shard = hash(user_id) % num_shards

SHARD 0

user_id: 0-24%

SHARD 1

user_id: 25-49%

SHARD 2

user_id: 50-74%

SHARD 3

user_id: 75-99%

Each shard = independent DB with ~25% of total data. Add more shards → more capacity. Challenge: Cross-shard queries and shard rebalancing.