Doubts about path that Nebulet is taking. #72
Comments
Nebulet doesn't have to be a general purpose kernel, it may not even be specialized for something, all it has to do is to offer the building blocks for its users to create a "specialized" os that does that one thing and it does well. |
So they are already doing that? there's also WASI that can be made to work on ButtABI, which then runs in my butt.
sound like a unikernel to me.
Why? Why would it be more efficient to compile some bytecode (wasm) just-in-time to native machine code? Or are you saying that there should be a nebulet "compiler"/"builder", that can compile one wasm executable ahead-of-time for a specified cpu architecture, so that we end up with a some-sort-of native unikernel? |
that's what i was thinking at the time |
I recently found this comment on hackernews about an envisioned platform that sounded suspiciously similar to nebulet, just having taken a different path early on.
Here's the comment in full:
- vinceguidry
It's food for thought. Should nebulet have taken a different path earlier on? Really, the essential concept here is architecture-independent processing of data. It doesn't matter if nebulet is a microkernel or what it's running on or what the drivers are written in.
It just need to be a platform that you can drop code on that will receive data, process it, and send it back out.
The text was updated successfully, but these errors were encountered: