On 4/27/07, Andrew Lentvorski <[EMAIL PROTECTED]> wrote:
Joshua Penix wrote:
> So one big question is: KVM or Xen? I was originally assuming Xen, but
> there's something to be said for the simplicity of KVM and the fact that
> it's right there in the mainline kernel. For those who don't know about
> it, here's a little reading that may be of interest:
>
> http://kvm.qumranet.com/kvmwiki (project home)
> http://tinyurl.com/3d5xad (good c|net article overview)
From:
http://www.gridvm.org/xen-vs-kvm.html
KVM "Uses QEMU for guest instantiation, but is not an emulator itself"
To my mind, that would simply rule out KVM.
QEMU just never gives me the feeling of having a good "forward vector".
The amount of time it has been taking to get QEMU to compile with
gcc4 has been huge. Their testing has also struck me as pretty haphazard.
Xen may or may not be technically better, but they at least seem to have
people driving the direction sufficiently to actually stand up and argue
with the VMWare people. The fact the RedHat is supporting it for CentOS
5 is probably a nice support vote, too.
So, I vote Xen on the strength of community. I can be swayed by
technical arguments, though.
Based on some experience with QEMU, I interpret the above statement as
"There exist QEMU support routines that can be used to turn a raw disk
image into a KVM image".
carl
--
carl lowenstein marine physical lab u.c. san diego
[EMAIL PROTECTED]
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-steer
