Support for TKG on VxRail was previously tested and documented in the Virtualizing GPUs for AI with VMware and NVIDIA Design Guide. This white paper provides more supporting information about how TKG can be used for containerized and virtual machine applications for our multi-ISV, multiapplication CV platform design.
Implementing this design expands the benefits of the platform to include the following:
- Access to a supported container management service using TKG on VxRail leveraging an existing validated design that customers have relied on.
- Support for NVIDIA
- GPUs with containers using the NVIDIA GPU Operator for TKG.
- Direct integration of TKG management into VMware vSphere can enable single-pane management of the ESXi clusters and services.
- Host virtual machines alongside containers in the same vSphere environment for complex environments leveraging any VMS and CV applications that are not container-based.
- Support for multiple worker/guest clusters and/or tenants instantiated on a common ESXi cluster managed with vSphere.
In summary, adding TKG support to the CV platform allows the ability to host a diverse range of both VM and container-based workloads on a common platform.