Messaging Queue

1. Clusteris collection of brokers. Inside brokermultiple topicsare present. Topic contains messages that producersare producing.

=> 1 broker: will have multiple topics
=> each topic: corresponds to a file 

Topic itself is a queue.

• All the messages that are stored inside topiccomes with a retention policy which basically states for how many days a particular message should be stored inside a topic. If the default retention policy is 7 days ,all the messages inside a particular topic will be basically stored for 7 days. After that the messages will be deleted from the topic. Internally also topics are basically implemented using file. You can open the file and you will see that data is stored inside binary format.

2. **Kafka Partitions** : sharding strategy in kafka.

3. Replication RF=2 means data will be replicated to 2 other nodes. Messages inside partition are basically stored ordered by time and are immutable and are stored in log format.