• Each kernel instance can run on dedicated CPU cores while sharing the underlying hardware resources.

    This sounds like a recipe for spending decades tracking down bugs and obscure race conditions. The kernel is simply too massive, and with too many vendor-specific workarounds, for them to feasibly catch and ensure compliance for every call to raw hardware.

    • There’s an underlying kernel under the kernels for each core that controls access to hardware. It has all the hardware drivers and maintains state.