Research topic
Kubernetes platforms
Kubernetes runs containerized workloads across a cluster of servers. For GPU work, those workloads include model serving and LLM inference, agents and the tools they call, search and retrieval, and scheduled batch jobs. Workloads ask for GPUs, and the scheduler places them on the nodes that have them. Namespaces keep teams apart on one cluster, network policies set the paths each workload may use, and admission policies reject workloads that break the rules before they start. With GitOps, the cluster follows a Git repository, so each change is reviewed and can be rolled back.
Research on Kubernetes platforms
The research here works from the Kubernetes documentation, project release notes, vendor documentation and published studies: how LLMs are deployed, scaled and rolled back on a cluster, how GPUs are scheduled onto nodes, how tenant isolation is tested, how GitOps keeps each model deployment reviewed and reversible, and how GPU node errors are detected and handled. Every figure carries its source and its date.
- MIG, MPS or time-slicing: sharing GPUs on KubernetesMIG vs time-slicing vs MPS on Kubernetes: what each GPU sharing method isolates, what measurements show, how the device plugin exposes it and how to size a slice.Kubernetes platforms7 min readPDF report
- Slurm or Kubernetes for GPU clusters: how to chooseSlurm vs Kubernetes for GPU clusters: the unit each scheduler places, how a training job starts all at once, where each fits and how one cluster runs both.Kubernetes platforms7 min readPDF report
- Deploying LLMs on Kubernetes: serving, scaling and rollbacksLLM deployment on Kubernetes, layer by layer: model servers, inference-aware routing, scaling on queues and token backlog, scale to zero and rollbacks by weight and revert.Kubernetes platforms7 min readPDF report
Get new research by email
Work with us on Kubernetes platforms
Rillor designs and builds Kubernetes platforms for superintelligence (SI) work on GPUs, on servers you control. Every change ships from Git, and each workload reaches only what it needs. The cluster sits on servers you own, on systems we supply, or inside a private SI cloud.
About Kubernetes platformsNotices
NVIDIA, AMD, Ceph, Kubernetes and related product names are trademarks of their owners and are used only to identify products. No affiliation or endorsement is implied.
Tell us what your cluster must run.
Send the workloads, the GPUs and servers they run on, the teams that deploy to the cluster and who runs it afterward.