Database sharding is a horizontal scaling technique where a large database is partitioned into smaller, independent segments called shards, each shard containing a subset of the total data and residing on a separate database server. Unlike vertical scaling (adding more CPU/RAM to one server) or replication (copying the same data to multiple servers for read scaling), sharding distributes write operations across multiple servers — enabling systems to handle write throughput that a single database instance cannot sustain. For WordPress, database sharding is an advanced architecture typically reserved for very high-traffic multi-site networks or applications processing millions of rows daily — the kind of scale seen at WordPress.com or large media networks, not typical business or content marketing sites. Standard WordPress does not support sharding natively; it requires a HyperDB or custom database abstraction layer. Most WordPress performance challenges at normal business scale are solved far earlier via caching (page cache, object cache) and query optimization, making sharding unnecessary until genuinely extreme traffic demands it.
Glossary entry