I am evaluating use of libvirt to manage a cluster of ESXi servers.

We plan to use "resource pool" concept of vcenter/vmware to easily scale-up or scale-down the cluster of ESXi servers

From vmware doc

"Resource pools are partitions of computing and memory resources from a single host or a cluster"

Since VM can be designated to hosts, clusters or "resource pool".. assigning VM to resource pool allows a simple way to manage scale-up / scale-down operations. 

This is because one doesn't need to know identity of ESXi server (from cluster of servers) to assign/place a VM.. one can just refer to resource pool.

From reading the doc of libvirt, it seems that libvirt always requires identity of ESXi servers in all the (vcenter) API calls. Is my understanding correct? 

If so, does libvirt have a similar concept/notion so that VM life-cycle can be managed without knowing identity of underlying server?

with regards

Ashutosh