Give me 2 mins, and I'll teach you how Kafka works:

• Kafka is a distributed streaming platform that stores & delivers data as ordered messages.

• Kafka often stores data on broker disks (SSDs) for durability.

• With diskless Kafka, it's possible to write data directly to object storage like S3.

• Broker disks then become only a buffer for writes & a cache for reads.

• This reduces costs and I/O bottlenecks, but reads and writes might be slightly slower.

• Plus, it keeps the Kafka guarantees, such as ordering, durability & transactions.

• An operator can choose per topic: fast (SSD) or cheap (object storage).

• It's best for data where cost matters more than latency.

(What else would you add?)

———

💾 Save this for later & restack to help others learn Apache Kafka.