Friday, Payday, Yay Day!
It's Friday, it's payday and as some others in the team like to call it, Yay day! What a long old week it has been. I won't go in to detail as I've moaned enough in my previous few posts. I'm actually going to have one of those old man sysadmin rants instead.
Kata Containers...

"as light and fast as containers", "also delivering the security advantages of VMs"
I was intrigued, so I looked at their "Learn" page and was presented with an overview.

Now compare this with "traditional" virtual machines:

So Kata Containers are basically re-creating the Type 2 Hypervisor but with Docker and Kubernetes features shoe-horned in for good measure. At this point, I'm struggling to figure out why you'd bother and not just go with a full-blown VM where creation, and images could already be automated. You start with a reference image or an ISO file, and then use something like Ansible/Puppet/Chef/etc. to get it to the state you need and keep it there without drift.
With KC, you're also limited to just Linux right now, whereas with a VM... crack on and use whatever you need.
Technology really feels like it's going around in circles and re-inventing things that already exist, except with more dependencies, a greater learning curve and more complexity for shits and giggles.
I've got some time off in a couple of weeks, so I'm going to try and spin a couple up and compare them to the resource usage of traditional VM's running on a bare metal (Proxmox) and a type 2 (Hyper-V) hypervisor to and compare the resource usage.
/FridayGrump