Re: Feedback on New Kafka Project

[Alviro Iskandar Setiawan <alviro.iskandar@gnuweeb.org>]

On Mon, July 20, 2025 at 10:11 AM Xiaorui Wang wrote:
> Hello Alviro
>
> I observed your interest in a number of open source projects, so I wanted to share a relevant open source project I'm working on.
>
> AutoMQ is a Kafka solution developed on object storage. This provides a number of key benefits:
>
> - 10x more cost-effective than Apache Kafka.
> - Partition reassignment 800x faster than Kafka.
> - 5x higher throughput in catch-up read.
>
> I'm excited to let you know that AutoMQ has successfully simplified the operation and maintenance of Apache Kafka for Grab.com, which has greatly lowered their cloud expenditures.
>
> Is this something that could be beneficial? I'd love to hear any feedback you might have!
>
> (If this isn't a good fit, don't worry – I won't reach out again!))
>
> Thanks,
> Xiaorui
> github.com/automq/automq

That sounds seriously impressive. It would be great if you can show a
comparison of your benchmark result of Apache Kafka vs AutoMQ.

Does it also utilize the recent shiny kernel feature, io_uring?

I think integrating io_uring could be a cool way to boost I/O
performance even further, especially with your network and object
storage setup. Curious to hear your thoughts on that.

io_uring will likely be the next big thing. We are still early...

-- 
Software Engineer & ITPM Officer
Alviro Iskandar Setiawan