Replication & Partitioning

A single database server has hard limits on storage, throughput, and availability. Replication and partitioning are the two fundamental techniques for pushing past those limits. Replication copies data to multiple nodes for availability and read performance. Partitioning splits data across nodes for write throughput and storage capacity.

These concepts come up in virtually every system design deep dive. When the interviewer asks "how does this scale?" or "what happens when this node goes down?", you're talking about replication and partitioning.