Vmware vs Virtualbox vs KVM vs XEN: virtual machines performance comparison

Written by Gionatan Danti on . Posted in Virtualization

User Rating:  / 167
PoorBest 

MySQL performance: Sysbench complex test

Now it is time to go ahead with some more heavy SQL test. The Sysbench complex is a test with, err, complex statement and mixed read and write requests. I run this test with 16 threads asking for a total of 10000 requests (1/10 that the simple test).

Some number in transactions per second:

Sysbench complex TPS

In the heavily threaded MySQL program, Xen is again the slowest machine by a very wide margin, while the others are more or less on par.

What about CPU load?

Sysbench complex CPU load

VMware is the “lightest” hypervisor, probably thanks to its paravirtualized network driver and low disk write CPU load (see the IOMeter results above).

Apropos of disks, the next graph is very interesting:

Sysbench complex HD load

KVM seems to be the hypervisor less heavy on disk subsystems. Don't be fooled by relatively good Xen results: its low load is probably only a results of its bad number of transaction per second.

UPDATE: a recent article comparing KVM vs VirtualBox can be found here: http://www.ilsistemista.net/index.php/virtualization/12-kvm-vs-virtualbox-40-on-rhel-6.html

Comments   

 
#1 Nathan 2012-09-12 03:12
This is a terrible review, to install the VMware paravirtual drivers but not the KVM Windows paravirtual drivers. All results from VMware must be discarded for comparison purposes.
 
 
#2 Marcelo 2015-11-15 03:16
A quick comparison I made between VMware Workstation Player and VirtualBox, with XP as guest, shows a ridiculous I/O advantage of VB, while VMware has a big advantage on 3D graphics.
 
 
#3 Gionatan Danti 2015-11-15 09:32
Quoting Marcelo:
A quick comparison I made between VMware Workstation Player and VirtualBox, with XP as guest, shows a ridiculous I/O advantage of VB, while VMware has a big advantage on 3D graphics.


Hi Marcelo,
VBox higher I/O speed probably is an artifact of VBox not honoring write barrier (synchronized writes) by default. While this give much higher speed, storage consistency is somewhat reduced and I do not suggest to disable write barriers on production host/machines.
 

You have no rights to post comments