AGUASTO

GPU resource orchestration

Kubernetes GPU orchestration — NVIDIA device plugin

Expose NVIDIA GPU devices as Kubernetes resources and allocate them to container workloads.

Original Aguasto conceptual illustration. Not manufacturer geometry, an operating Aguasto asset, or a measured benchmark.

When it is relevant

  • Schedule GPU-backed containers on an existing Kubernetes cluster
  • Study device allocation, GPU feature labels or documented sharing modes
NVIDIA Kubernetes device plugin official repository (external source) · Kubernetes official GPU scheduling guide (external source)

Hardware and software assumptions

Documented ecosystem: NVIDIA. Platform support is not a guarantee that every checkpoint, optimization or feature works on every device.

  • Install compatible NVIDIA drivers, container runtime tooling and the device plugin on GPU nodes
  • Check cluster-version and plugin-version compatibility
  • Supply a separate model-serving engine and operational monitoring
  • No Aguasto Kubernetes cluster or GPU inventory is asserted
NVIDIA Kubernetes device plugin official repository (external source) · Kubernetes official GPU scheduling guide (external source)

Operating trade-offs

  • This precise plugin profile is NVIDIA-specific; other vendors require their own plugins
  • Device allocation does not tune inference performance or provide a complete GPU operations system
  • Time-slicing shares GPU memory and fault domains and is not equivalent to isolated hardware partitions
NVIDIA Kubernetes device plugin official repository (external source) · Kubernetes official GPU scheduling guide (external source)

Continue the research

Choosing an inference frameworkPrefill, decoding and disaggregationBenchmark methodologyModel referencesInfrastructure collaboration

Sources & verification

Reference revision 1 · Reviewed . Manufacturer claims and documentation are attributed; Aguasto has not run the cited products in production.

  1. NVIDIA Kubernetes device plugin official repository ↗

    NVIDIA device plugin and GPU feature discovery; driver and runtime prerequisites; time-slicing memory and fault-domain limits; distinct sharing modes. Publication: date not stated in the reviewed source. Verified: 2026-10-12.

  2. Kubernetes official GPU scheduling guide ↗

    vendor drivers and device plugins; GPU extended resources such as nvidia.com/gpu; resource requests / limits rules; node-selection constraints. Last modified: 2024-09-20. Verified: 2026-10-12.