In Kafka, topics are divided into parts which are called partitions. 

Partitions allow one or more consumers to read data from servers in parallel. Read and write responsibility for one particular partition is managed on one server, called the leader for that partition. A cluster may have zero or more followers in which replicas of the data will be created. Replicas are merely copies of the data in a particular partition. The followers do not have to read or write the partitions separately; rather, they just copy the leader.

Partitions in Kafka are used to increase throughput. Replicas ensure fault tolerance.

https://developer.confluent.io/learn-kafka/architecture/geo-replication/