Yes, node pool is a GKE-specific concept. It allows you add groups of nodes into the cluster that have different resources or characteristics (memory, cpu, disks, preemptible, etc).

Kubernetes itself is only aware of nodes registered in the cluster, their resources, labels and taints. With labels, you can schedule workloads to specific node pools using Node Selectors. Taints allow you repel all workloads that don't tolerate the properties. See Taints and Tolerations