From Metal to Model.. building a GPU Cloud in a Box.
Mirantis presented at AI Infrastructure Field Day 3 about supporting public, private, and hybrid cloud AI integrations. How does an organization go from Public, Private, and Hybrid Cloud to an AI Cloud?
Mirantis talked about how it means changing the model. They are bridging the enterprise data center like how cloud has change it. You choose your model and Mirantis can provide solutions to match.
If you’re not sure what OpEx model you want to use, Mirantis gives you flexibility to move between OpEx and CapEx as you figure it out.
Mirantis doesn’t require you to build out a big footprint inside your data center, just like the cloud migration. You just need a Kubernetes. K0rdent is the control plane on top of Kubernetes. If you have basic Kubernetes knowledge, you can figure out how to use Mirantis, through the extraction of complexity of YAML.
So, where does Mirantis fit compared to say Rafay?
Mirantis says their fit for enterprises is because of their Dynamic Control Plane. Like the cloud took complexity out of the on-premise data center, Mirantis is taking away that complex control plane. The whole process is dynamic. AI has been single tenant focused, Mirantis is changing it to be Multi-Tenant control plane.
Less hardware and the ability to dynamically control and reduce the complexity is changing the Metal to Model load for AI workloads.
Check out all the coverage and details from AI Infrastructure Field Day 3.