[PATCH] Add a VirtualBox section to Handbook's 'Virtualization' chapter

[Simon L. Nielsen]

On 2009.07.19 11:34:38 -0400, Tom Rhodes wrote:
> On Sun, 19 Jul 2009 11:02:59 +0300
> Manolis Kiagias <sonicy at otenet.gr> wrote:
> 
> > Marc G. Fournier wrote:
> > > On Sat, 18 Jul 2009, Manolis Kiagias wrote:
> > >
> > >> But isn't qemu distinctly different in the way it works from a
> > >> virtualization program like VMWare or VirtualBox? I think the first
> > >> paragraph serves well as a gentle introduction to the topic.
> > >
> > > How different?  Wine, I could see, but qemu?  Please elaborate ...
> > 
> > AFAIK qemu also provides processor emulation, thus is mostly referred to
> > as an emulator rather than a virtual machine. I am no expert on this
> > though - I've used qemu in the past but could never get the level of
> > performance possible with VirtualBox or VMWare (which beats them both I
> > believe).
> 
> Interesting - see, I tried doing an install of qemu a long
> while ago but never went beyond installing it.  Though, from
> what I have read, I would not consider it a "virtual machine
> host solution" either.  While I understand it runs image files,
> I don't think it's geared for several OS images, running
> concurrently.  Again, note, I only installed - and when the
> image I had (passed by a friend) failed to run, I just removed
> it.

I really don't see the big difference between qemu and VMware /
Virtual Box.  qemu is more flexible in hardware support - VMware (and
Virtual Box?) is faster.

Just like VMware server multiple concurrent virtual machines just run
in different processes.  Yes, qemu is slower than VMware but e.g. for
testing og kernel hacking it works nicely.  I can't comment on Virtual
Box's speed as I haven't tried it.

[reordered]

> > The paragraph was loosely based on the original one stating "No
> > virtualization solution for FreeBSD as a host". Seems the original
> > author also did not consider qemu as a virtual machine in this sense.

Yes, it came from there, it wasn't any more correct IMO there :-).

I guess my basic problem with the paragraph is that it seems to me to
be praising Sun a tad much for releasing Virtual Box while ignoring
that qemu has been available freely for years, but perhaps that's just
me.

Anyway, I think I made my point and I will let it be up to Manolis as
the author was should be in the section and what should not.

-- 
Simon L. Nielsen