A mechanism for distributing data across multiple servers.
A mechanism that horizontally splits data across multiple nodes based on something like a partition key, allowing scaling beyond the limits of a single node. Adding or removing nodes after the system is already running requires redistributing existing data (resharding), which carries a real risk of degraded performance or downtime while it's underway -- so getting the initial sharding design right matters a lot.
© 2026 ITBGM