Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The technical reason is the inherently distributed and shared nature of Kafka setups makes writing an ACID-compliant frontend (basically?) impossible.

If your application assumes ACID compliance, it will fail. Also, I haven’t looked into this one as directly, but I imagine that query latency compared to an RDBMS would be quite poor.

IMO, most people reach for Kafka too soon, where an RDBMS with a proper schema that includes transaction and valid history would better suit the domain. It’s possible to write very efficient and ergonomic “append-only”, temporal data schemas in Postgres schemas.



But there are a lot of usecases where ACID is not needed and we can do with eventually consistent states. Btw, you do get C and D.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: