Kafka Tiered Storage Architecture
Data & AI Pipelines diagram · made in Skemio

What it shows
- Active log segments remain on local broker NVMe drives for low latency.
- Sealed log segments upload asynchronously to object storage like S3.
- Brokers fetch historical segments transparently for lagging consumers.
The takeaway
Tiered storage decouples compute from storage, turning Kafka into a long-term replay log.
- #ApacheKafka
- #Kafka
- #DataEngineering
- #SystemDesign
- #DistributedSystems





