Miluba made a comment in YouTube about Kubernetes on metal with net boot possibly reducing power bills. I hadn’t really thought of it, but this could significantly reduce costs for the enterprise by allowing certain nodes to “sleep” (we’ll say) when not in use. Kubernetes could maintain a buffer of active systems so that there would never be any lag, but imagine the cost savings to allow all of that hardware to be managed and scaled as if it were pods.
There’s a good chance this won’t come from the cloud providers because they love all the idle nodes running and burning up billable CPU. But Amazon, at least, championed elastic infrastructure as a marketing method and maybe would see the value of K8S-on-Metal to allow for this, but since everything on Amazon is already virtual maybe not. It is going to take more enterprises to realize that running K8S on virtual servers is 30% less efficient to create the market pressure to provide a bare, on-metal solution.
This is all very intriguing and tends to make me believe more strongly that K8S-on-metal is the future (at least from where I’m sitting).
Tags:
#k8s #cloud #talos #sidero #metal #devops #infrastructure