Replication 在多个服务器上维护数据副本——用于 可用性(抵御故障)、读扩展(从副本提供读取)和 地理分布。不同的策略(primary-replica、multi-primary、同步 vs 异步)涉及权衡。
为什么需要复制
✓ AVAILABILITY/redundancy → if one copy fails, others have the data (no data loss, failover)
✓ READ SCALING → serve reads from replicas (offload the primary) → handle more reads
✓ GEO-DISTRIBUTION → copies near users (lower latency); disaster recovery
→ replication is fundamental to reliable, scalable data systems.
