Cassandra stores copies (called replicas) of each row based on the row key. The replication factor refers to the number of nodes that will act as copies (replicas) of each row of data.