i am having a server with 16 virtual core and 24G RAM,using Xen virtualization and ubuntu as dom0 Created 4 VMs (in para mode),each with different applications.

CPU Load vary on each vm,somtimes first vm reaches nearly 100% CPU and others under 25% or even less.

So is there a way in which vm can get cpu from other vms when they are not actually using it or utilization is under 25%.Same in the case of RAM also.

I am not sure whether i am mentioning Cloud here.

Initially i would like to give every vm a single VCPU,but can scale up to 8 or more by taking cpu from other vms if they are not using it.

Is there any kind of tool that makes vm to scale its resources when demand increases.

Is cloudstack and openstack designed for these kind of purpose or is that just a GUI to manage VMs.