Your team has an existing GKE standard cluster running web applications. A new data science team wants to deploy machine learning workloads to the same cluster, but their containers require GPUs. The current cluster nodes do not have GPUs attached.

What is the BEST way to accommodate this new workload?

A GKE cluster can contain multiple 'node pools'. A node pool is a subset of node instances within a cluster that all have the same configuration (machine type, disk size, GPUs). To add GPU capabilities to an existing cluster without disrupting current workloads, you simply create a new node pool configured with GPUs. You can then use Kubernetes node selectors or taints/tolerations to ensure the ML workloads run only on the GPU nodes.