openwhisk-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rodric Rabbah <>
Subject Re: Add support for microkernels instead of containers
Date Tue, 17 Jul 2018 21:23:54 GMT
Hi James

There’s an abstract interface to the execution unit in the invoker: Start/Pause/Resume/Stop/Logs.
You can select the implementation through a configuration deployment (SPI). 

There was some work on using the interface I alluded to for unikernels. I’d imagine the
interface it can be adapted for working with a process, microkernel, ...

I’m suggesting it’s a drop in replacement but at face value I don’t see that it’s
necessary to be too invasive. The openwhisk core is really about starting/pausing/resuming/stopping
an execution unit (which happens to be a container today). 

I think containers are too coarse grained an execution unit for functions and expect technology
to change in the future. But when/how long it will take... we’ll see. I’m curious to see
how your work unfolds with OpenWhisk and we’re happy to help. 


> On Jul 17, 2018, at 1:10 PM, Farwell, James C <> wrote:
> Markus Thoemmes found my question on StackOverflow, and suggested I post it here;
> I am trying to customize OpenWhisk to call a microkernel from the Invoker, rather than
Docker. Is there an effort underway currently to add this support, or a development guide
covering the changes I would need to make? My current understanding of the code is that this
will be a substantial project.
> Is there guidance available on how to move away from the concept of containers? Or will
I be better off treating a microkernel as an abstracted type of container?
> Thanks,
> --James

View raw message