XCDR is a powerful feature that allows data replication between different-sized clusters, making it an excellent option for disaster and recovery plans, as well as bringing data closer to users. The replication process is memory-to-memory, with writes saved in memory first and then sent over the network through multiple threads, limited only by network speed. XCDR is topology-aware, automatically re-establishing connections when nodes are added or removed from the source cluster. It also provides rack zone awareness, protecting against multi-node failure events by separating active data and replicas across groups mapped to different racks, zones, or VM hosts. Replications can be configured as unidirectional, bidirectional, or hybrid, with pause and resume capabilities via Database Change Protocol (DCP). XCDR relies on DCP to propagate changes, guaranteeing the same document is replicated among all clusters regardless of connectivity problems. Conflict resolution and data filtering are also key features, allowing for consistent selection of document versions and filtering of replication data using advanced language constructs. Overall, XCDR offers a simple yet powerful way to manage data across multiple clusters, with minimal maintenance required.