[olug] VMWare Player vs VirtualBox

Dan Linder dan at linder.org
Tue Nov 23 17:29:39 UTC 2010


I've been a long-time user of VMWare Workstation and Player, and have used
VirtualBox a bit, but never really compared them back-to-back.  Since I was
needing to setup some testing systems, I decided to give VirtualBox a try.
 Just some background, my host system is a Ubuntu 10.04 system, 12GB RAM,
and an AMD Phenom II X4 cpu.

For the tests, I installed the latest version of both:
 * VMware Player version 3.1.3 - http://www.vmware.com/vmplayer
 * VirtualBox version 3.2.10 - http://www.virtualbox.org/

<http://www.virtualbox.org/>Both installed easily, but both were a manual
process.  The Ubuntu 10.04 repository only has VirtualBox 3.2.8 so I had to
find notes on installing the latest 3.2.10 version.  VMware is a single
download of a 99MB "bundle" file (actually a shell script), and running it
as root got it up and running.

Functionally both achieve the same thing and allow you to run multiple guest
OS's in windows on your host system.  You can configure how much RAM, number
of CPUs, etc each virtual machine see, then load the OS of choice onto them
and start running.  I'll let you do the actual point-by-point comparison of
the supported OS's and features - in short, most users will find that both
achieve the same goal.

Both are free to use for personal use.

I have these comments:
 * Under CentOS 5.5 x64, when I open a terminal window and hold down on a
key too long it appears to be stuck down and nothing will get it un-stuck.
 It's so locked up that the screen refresh seems to have monopolized the GUI
enough I can't use the mouse to kill the terminal window.  And speaking of
the GUI...
 * Again, under CentOS 5.5 x64, things that involve moving/resizing windows
seems sluggish.  As a test, I ran "glxgears -info" on both systems and the
VirtualBox was anywhere from 600fps to 800fps, where the same test under the
VMWare Player guest ran at a solid 2200fps.  This may be due to the fact
that the 2D acceleration is only available on Windows systems in VirtualBox
at the time.
 * Using "top" to check the CPU usage of both virtual systems, VirtualBox
always kept 100% CPU (core) in use, where VMWare would go down to around
25-35% when just sitting at the desktop.
 * In VirtualBox's defense, there were times that it seemed a little bit
snappier in some instances, but this was lost in the video issues mentioned
above.  Maybe if you're running a CPU intensive tool in a text-only console
system, VirtualBox may have an edge.
 * A drawback to VMWare is that it's not easy to setup a "server" to
automatically start a series of virtual machines, nor can you pause or shut
one down from the command line.  (Please correct me if I'm wrong.)

I know some people just hate the "VMWare corporate image" - if you're one of
them, then VirtualBox is your choice...er, um, wait, VirtualBox is owned by
Oracle now that they've bought Sun.  That's a "Rock & A Hard Place" decision
for some...

Conclusion?  I'm going to stick with VMWare Player for my personal
workstation needs.

Just my $0.02 cents.

Dan

-- 
***************** ************* *********** ******* ***** *** **
"Quis custodiet ipsos custodes?"
    (Who can watch the watchmen?)
    -- from the Satires of Juvenal
"I do not fear computers, I fear the lack of them."
    -- Isaac Asimov (Author)
** *** ***** ******* *********** ************* *****************



More information about the OLUG mailing list