What is Cassandra's coordinator node?

A coordinator node in Cassandra is whichever node receives a client request — any node can be a coordinator since all nodes are equal (no master). The coordinator's responsibilities: Route the request: hash the partition key to determine which nodes hold the replicas. Contact replicas: send the read/write to the appropriate replica nodes. Handle consistency: wait for responses from the number of replicas required by the consistency level. Return result to client: aggregate replica responses (for reads: select the most recent data; for writes: confirm success). Handle failures: if a replica is down, route to the next replica in the ring. Write to a hinted handoff if needed. Drivers: modern Cassandra drivers (DataStax Java/Python/Node drivers) are token-aware — they calculate which node is the correct coordinator for each query and send directly to it, eliminating an extra hop. This reduces latency and coordinator load. In production, balance coordinator load across nodes using a token-aware driver and multiple contact points.