What model does Kafka use?

Apache Kafka uses a publish-subscribe model with a distributed message log. Producers publish messages to topics, which are divided into partitions. Consumers read messages from these partitions, with each consumer able to read independently, ensuring scalability and fault tolerance.

Key features of Kafka's model:

• Messages are ordered within partitions.
• Consumers can read messages from any offset.
• Supports both queue model (one consumer reads a partition) and publish-subscribe model (multiple consumers read from different groups).

Thus, Kafka combines the message queue model and the publish-subscribe model with scalability and high performance.