It's obligatory for me to point out that the same engineers built a proof-of-concept component for Diego[0], the scheduler/manager system in Cloud Foundry. This is using the same pathway that allows Cloud Foundry to support buildpacks, Docker images and .NET apps under the same scheduler.
So if there was enough demand, this is a viable pathway to having unikernel software managed by a fully mature cloud platform. Personally I think that the OSv/Java unikernel stuff is probably in the sweet spot of industrial demand and plausibility, since programmers on the JVM are accustomed to operating at a fair distance from the OS's APIs.
Disclaimer: I work for Pivotal, we donate the majority of engineering to Cloud Foundry.
Could you email me? I've beat the drum a few times on us supporting unikernel workloads, it'd be nice to have an existing case to suss out what's hard and what's easy.
I'd also be very interested to know more about your use of unikernels and the tooling around them. Would you mind getting in touch? Email is in my profile.
This looks like a great project! Although the name pronounced in my head (for me at least) looks like "Eunuch" so slightly unfortunate (I see on the page it is pronounced unique, but just thought I would point that out).
I suggested you will watch my session (https://goo.gl/7Tm17A) - I am focusing on the use cases, vision and the motivation of project UniK. you can also you can read my blog (https://goo.gl/bZVoLr) to get more context.
hope that help.