I am setting up private cloud for some experiments using xen as the hosting system. But I am faced with a problem for which I can't seem to get solutions.
I have to do some kind of automatic provisioning of VMs given the server load. Eg: if server of type A gets to lets say 60% load the cloud should spawn off another vm instance of the same type to distribute the load(using the netscalar).
Is there an opensource system that can help me or how do I go about developing scripts to do the same.
If I understand you correctly, you want to live-migrate the VMs depending on the load of the host. You can use OpenNebula to help you with this. You can use the advanced scheduler named Haizea with OpenNebula. While I've never tried this, but you can use this with ONE's APIs to create more VMs if a VM gets too much load.