How does partitioning differ from sharding?

Partitioning and sharding share the same idea, in both cases large data gets split into parts. But the key difference is where and why those parts are stored.

#Partitioning

• Where: every partition lives on a single server (within one database).
• Goal: improve performance and manageability within a single system.
• Management: handled by the DBMS itself (e.g. PostgreSQL can partition tables natively).
• Typical scenarios: faster queries, cleaning up old data, optimizing storage for large tables.

Example: an orders table split by month, January, February, March, all stored on one server.

#Sharding

• Where: each shard lives on a separate server or cluster.
• Goal: horizontal scaling of the whole system and load distribution.
• Management: requires external logic, request routing, synchronization, replication.
• Typical scenarios: large web services, millions of users, terabytes of data.

Example: users with IDs 1-1,000,000 on one server, the rest on another.

In short:

• Partitioning = splitting a table within one server.
• Sharding = splitting a database across different servers.