How do sharding and replication work in databases for scaling?

Replication is the process of copying data from one database server to one or more other servers. Typically, a primary node is designated for writes and replica nodes for reads. This approach allows scaling read performance and ensures high availability of the system in the event of the main server crashing.

Sharding is a method of horizontally partitioning a large database into separate parts, called shards, which are placed on different physical servers. Each shard contains a subset of the total data volume, making it possible to distribute both read and write loads among independent nodes.

Unlike replication, where each node stores a full copy of the data, sharding solves the problem of physical limits of a single server in terms of disk space and computing power. However, implementing sharding significantly complicates the application architecture and the writing of queries that require combining data from different shards.

Before implementing these architectural approaches, it is necessary to analyze the following aspects: