unik VS linuxkit

And on that note, I just found this list of UniKernel projects:

http://unikernel.org/projects/

I have especially had hopes for the UniK [1] project, as it was/is written in Go AFAIK. I see now it incorporates work from the Mirage project as well. Not sure what is the status of this project anymore though.

[1] https://github.com/solo-io/unik

Wow, just now seeing this topic. I work for a cloud company hosted in AWS. We started out, Netflix/Spotify style microservices. We were all on ec2 images generate by packer (and later with AWS Image Factory). When Docker hit, we kicked the tires but never did anything with it beyond using it for running unit tests, and later, infrastructure tests. 5 years ago, during a hackathon, our little group began experimenting with Unikernels, or library operating systems. Interestingly enough, these Unikernels were all stripped down BSD kernels. OSv is FreeBSD based, and Rumprun is NetBSD based. Services running in EC2 on Unikernels would spin up and start sending and receiving traffic before the AWS EC2 healthchecks completed. They are blazing fast! Only problem in 2017, was the tooling. It would have taken too much effort to use Unikernals with our infrastructure. As soon as they start making Unikernels that can run Java bytecode like native code, the fate of containerization will be sealed, IMO. We could get basic JVM webservers running on OSv, but not Cassandra, not Kafka, not yet. OSv now runs on Firecracker, but I have not tried it out, yet. Some links if you are interested: OSv: https://osv.io Rumprun: https://github.com/rumpkernel/rumprun We used this tooling during the Hackathon, but doesn't look like it has been touched in 3 years: https://github.com/solo-io/unik Unikraft Unikernel Dev kit: https://unikraft.org/ And don't forget Firecracker running in Kubernetes https://www.weave.works/oss/firekube/ And of course, being a FreeBSD subreddit, let's not forget FreeBSD on Firecracker https://www.daemonology.net/blog/2022-10-18-FreeBSD-Firecracker.html

Crazy idea that I'm sure isn't an original thought: instead of adapting the languages to deal with abstracting the idiosyncrasies of each OS, change the OSes to expose a universal API to make everything else lighter.

I guess that's also kinda Docker or QEMU or V8, but also https://github.com/solo-io/unik if you think about it differently.

In other words: hey, Lisp Machines were an excellent idea back then, but they still are. Maybe someday we'll have a V8 co-processor. More fun reading: https://lobste.rs/s/2poahh/what_i_could_not_undiscover_about

The HN conversations around unikernels suggest that they're not ready for production yet [0] but feel free to set that record straight.

In the meantime, a handful of organisations/individuals seem to be working on becoming "Docker for unikernels". That's probably an unfair description, but they're aiming to produce tools for building and managing unikernels: Unikraft [1], NanoVMs/Nanos [2], Unik [3]. Other orgs are producing unikernel-based OSs and VMs [4].

What is your toolset for building and managing unikernels? What have you learned?

Bonus question: is Unik dead? [5]

[0] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&query=unikernel&sort=byPopularity&type=story

[1] https://unikraft.org/

[2] https://github.com/nanovms/nanos

[3] https://github.com/solo-io/unik/

[4] http://unikernel.org/projects/

[5] https://github.com/solo-io/unik/issues/172

UniK will compile and deploy its own 30 MB unikernel. This unikernel is the Unik Instance Listener. The Instance Listener uses udp broadcast to detect (the IP address) and bootstrap instances running on Virtualbox.

Another project that aims to deliver this is Linuxkit (https://github.com/linuxkit/linuxkit). All the components they ship are written in memory safe languages (usually Go) and run as containers under containerd. You can build a custom image very easily, fully defined as a YAML file.

docker-the-company maintained https://github.com/linuxkit/linuxkit when I worked there. I have no idea who maintains it now, but it looks like it is still active (presumably still docker-the-company, since their adopters list [1] lists docker desktop).

[1]: https://github.com/linuxkit/linuxkit/blob/master/ADOPTERS.md

LF-Edge EVE project leverages Linuxkit to create custom OSs for Edge Devices which in turn leverages Containers as Lego Blocks

LinuxKit - Docker

Behind the scenes Docker Desktop for Mac spawns a linuxkit VM with a bit of extra stuff like NFS to enable mounting Mac paths into containers. In the Docker Desktop settings you'll find the current resource assignment for that VM. That is pretty much reserved for docker so that it does not have to compete with MacOS processes for available resources.

That said... you might want to check out linuxkit

The definition of what a unikernel is needs to be narrowed down, a lot of these projects in the space (not all the ones listed above) have material differences that are not clear:

- some run only one language

- some require recompilation

- some essentially swap out libraries, others do something closer to dropping your already mostly static binary in a minimal disk image

- some build pid1 processes, others VMs images

Anyway, here are some additional entries in the space:

- https://ssrg-vt.github.io/hermitux/

- https://github.com/linuxkit/linuxkit (more embedded/minimal VM than unikernel)

- https://nabla-containers.github.io/ (runs on Solo5)

I am going through using Linuxkit to build AMIs for cloud providers now. I wouldn’t necessarily class linuxkit as a universal project because it doesn’t have the hallmark blurring of user and kernel space or kernel-as-a-library but you can customize the kernel so it’s an adjacent idea, and I think it’s the one most likely to be in actual use at non-hyperscalers.

I believe there is growing interest in providing leaner, "trimmed" runtimes for services deployed to the cloud. Today, this is seen largely by specializing the Linux kernel for, for example, container services[0] or in general[1], as much as that is possible (the paper above covers this problem in greater detail). But, Unikernels in themselves are not yet widely adopted. This is the space Unikraft is aiming to enter, providing the ultimate level of specialization for a target application.

It's clear that bigger players, such as Red Hat[2] are interested in the topic of unikernels, and that cloud providers are preparing for this future too [3].

[0]: https://github.com/linuxkit/linuxkit