Cluster (replication cluster): It consists of multiple nodes, and each node is redundant with the same function; in order to cope with the large amount of data that cannot be stored at a single point, and the high concurrency that cannot be processed by a single point in time.
Master-slave: The purpose is that the master writes and the slave reads. So the master-slave is just a node. For example sentinel mode .
Sharding (sharded cluster): When doing a cluster, the data of each node is the same, so the total capacity is limited by the minimum capacity of a single node. So the data can be split into multiple redis instances.