Comment by kentonv

2 months ago

This is 100% open source and self-hostable.

https://github.com/cloudflare/cloudflare-os

It runs on top of our open source runtime.

https://github.com/cloudflare/workerd

Does workerd have full feature-parity with Workers, including stuff like Durable Objects?

  • It's the same code. The only thing it doesn't include is the global scheduling and orchestration, but you don't actually want that locally. (And honestly... that stuff is a bit of a beast.)

    Durable Objects are fully supported by workerd (and Cloudflare OS uses them extensively).

    There is, however, one catch currently: Durable Objects don't scale out well without the global scheduling. For running Cloudflare OS for one user, this is actually no big deal, but a company-wide instance might not work well.

    But I'm actually fixing that:

    https://github.com/cloudflare/workerd/pull/6780

    Unfortunately didn't get enough time to land it before this release, but it'll come soon.

    With that I would expect a self-hosted Cloudflare OS on your own cluster to work quite well.

    • This is exciting! I've loved the idea of Durable Objects but I've disliked that there's no existing way to use them while scaling to multiple machines outside of Cloudflare.

It's cool that it's open-source, but how much effort and scale would it take to run it on non cloud flare infra?

  • ffs there is literally no pleasing some people. "Hosting an open source project is too difficult! They should provide a free SaaS solution instead on their own cost but also no lock-in!!"

    • I think it’s a legitimate question! Some companies operate at such a scale that if you tried to adopt their tools or services, it would be way too complex for the ROI.

      Think Bazel or how K8s was for ~ever. For Google’s scale? Required. For a small team? A net-negative.

      1 reply →

    • Adopting a new technology often needs a lot of consideration. People will have a lot of questions, want to understand the implications and discuss with the hive mind. No need to attack them for that, I think.

      1 reply →

    • High switching costs is a form of lock-in "open source" or not.. This is something people with real experience making real decisions are well aware of..

    • Open Source is a marketing term if your solution takes $2m worth of compute or devops time in order to run properly. The idea is, you have a way out if you want to leave. An option. But if leaving costs $x,xxx,xxx then that option was never there for most people. One example is NextJS. It's open source but even CloudFlare couldn't figure out how to run it properly on their platform.

      2 replies →

Kenton, back when I was building qbix.com/ecosystem you were building sandstorm, and I wanted to get in touch.

Now I have spent about 6 months building the hardened, self-hosted AI version that can be hosted in different clouds like AWS and GCP, and you shipped yours.

I think we’d have a lot to discuss. Should I reach out on gmail?