Confluent Developer ft. Tim Berglund, Adi Polak & Viktor Gamov cover image

Improving Apache Kafka Scalability and Elasticity with Tiered Storage

Confluent Developer ft. Tim Berglund, Adi Polak & Viktor Gamov

00:00

J-Troid: How to Access Data Transparently

The idea is that you split a topic up into the hot set. Once it's flushed from the page cache into disk, it's eligible for tiering. And then after that point, we would tier that eligible data to the remote object store. So instead of persisting on the desk, it would end up persisting inRemote Object Store (ROS) Something like S3 or DCP. Right? How do you manage replication when a node completely dies and has to be brought back up? That's what you're talking about. The first question is how do we still access the data transparently? We like nothing really changes at the user level. Nothing really changes for the user

Play episode from 03:58
Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app