Incus Everywhere

Incus
1

Hi Team,

I am starting to realize that I believe in an ‘Incus Everywhere’ vision of the world. I am writing this topic to help me vet the vision. This is a follow up to a previous post: Incus + Netbird + Phoenix => Good Synergy

The principle benefit is: the more places you run Incus, the fewer number of experts you need to hire. You use the same expertise everywhere.

Incus in the cloud:
Running Incus on a cloud bare metal server is quick, easy, stable, secure, and it can be completely managed from a script/cli/api perspective (no annoying webuis). Bare metal clouds are becoming more prevalent, and they represent significant cost savings over tradition AWS-like platforms.

Incus in the office:
Creating a dual-nic Incus instance to host office support services (firewall, storage, apps) is relatively simple and scalable. When you have 10+ offices, the simplicity of this solution is quite attractive. The fact that Incus plays well with openTofu (and similar tools) makes this statement even more compelling.

Incus on the desktop:
I have run Incus on my desktop since before it was Incus. That fact that I can use the same tool everywhere for all my development/deployment needs is quite convenient (and amazing).

My question to you (the community): is there anything else that stands out to you? Would you state any of these points differently. I appreciate your time and consideration. I hope posts like these are deemed helpful and positive. I appreciate this community!

Chuck

2

It’s definitely the goal of Incus to provide a near identical experience whether you’re running it on your laptop or in a large production cluster. That’s what really simplifies operation and its use in general.

That’s also why we’re spending time on things like Incus OS to standardize some of those bits even more, basically so that for most folks, you won’t even need to really care about the Linux OS management side of things.

You obviously still will on your laptop as that won’t be running Incus OS, but for any other deployments, whether it’s a home lab, office environment or datacenter, you’ll be able to use the exact same OS image and have the exact same behavior everywhere.

That may not necessarily fit everyone, there are obviously folks who want Incus to sit alongside other things on their servers, for those folks, using the same Incus packages as you’d use on a laptop, but possibly using Incus Deploy to deploy and manage those systems will still be supported and makes that pretty easy.

We’re also now looking ahead at what to do once we do have Incus OS ready for prime time.
The next obvious piece we’ll be focusing on is on deploying internal services on top of that Incus deployment (whether standalone or cluster). That means deploying things like a load-balancer for the API, the support services for Ceph, OVN or Linstor, an authentication/authorization stack, a monitoring stack, …

Our current plan around that is to make use of Incus’ OCI support so we can just have a bunch of simple, pre-tweaked, pre-tested application container images for those services, then have logic in Incus OS to automatically create an internal project and deploy the needed services to offer a fully functional production environment.

That’s basically the next step in ensuring that basically all Incus deployments will behave the same way, now expanding that to storage/network operation, authentication/authorization and monitoring.

3 Likes