Hi, On Mon, Sep 14, 2009 at 11:18, Dennis J. <dennisml at conversis.de> wrote: > On 09/14/2009 04:53 PM, Akemi Yagi wrote: >> On Mon, Sep 14, 2009 at 7:24 AM, Hildebrand, Nils, 232 <Nils.Hildebrand at bamf.bund.de> wrote: >>> KVM uses a para-virtualized approach? >> >> Not at this moment according to this Red Hat virtualization guide: > > Ugh, I guess that means my plans to switch from Xen to KVM have to wait > until RHEL 6 is released. I don't believe KVM will *ever* support para-virtualization in the same sense that Xen does. For instance, see this FAQ in KVM's website: http://www.linux-kvm.org/page/FAQ#What_is_the_difference_between_KVM_and_Xen.3F I believe the point is that current support for VT in recent processors is good enough to be able to run VMs with a native kernel at the same speed that could only be achieved with a para-virtualized kernel before. Therefore, the para-virtualized approach is being discontinued as "a hack" and the tendency is to improve VM technologies to run native code only. On the other hand, there is now talk about para-virtualized device drivers, which mean drivers that are optimized to run in a VM environment, which I believe are important in getting good performance from native kernels in VMs. The same concept exists in Xen, when you run Windows in Xen you do it using HVM (non-para-virtualized) mode, in which case you will only get good performance by loading the Xen drivers on the Windows machine, I believe the concept is the same there. HTH, Filipe