In Kafka, data is stored across different brokers. A particular broker may stop working at some point in time, which will eventually lead to the loss of data stored in that particular broker. To ensure that data is secure and not lost whenever a broker stops working, topics are replicated and stored across different brokers. So, if one broker stops working, the data stored in that broker can still be fetched from some other broker.
Kafka uses topic replication to ensure fault tolerance. In order to create a topic with multiple copies, the replication factor needs to be specified for the topic at the time of its creation. The minimum value of the replication factor is one. Its maximum value depends on the number of brokers present in the Kafka cluster. The replication factor cannot be greater than the number of brokers in the cluster.