Replication in Kafka is the process of maintaining multiple copies of each partition across different brokers so that data remains available and durable even if a broker fails.
Key Points: • Each partition has a designated leader broker that handles all reads and writes, and one or more follower brokers that replicate the leader's data. • The replication factor for a topic determines how many total copies of each partition exist across the cluster. • If a broker fails, Kafka can promote one of the surviving in-sync replicas to leader, so the topic stays available. • Replication also spreads read load in the sense that failover keeps consumers served, though by default only the leader serves client traffic unless follower fetching is explicitly enabled. • Higher replication factors improve durability and availability at the cost of extra storage and network usage.
Example: A topic with replication factor 3 stores three copies of every partition across three different brokers, so the cluster can survive up to two broker failures without losing that topic's data.
Interview Tip: A concise interview answer is:
"Replication keeps multiple copies of each partition on different brokers so a broker failure doesn't mean data loss or downtime — Kafka just promotes one of the in-sync replicas to leader — and the replication factor is the main knob for balancing durability against storage cost."