On Fri, 25 May 2001 21:08:01 +0800, John Summerfield wrote:

>> I see absolutely no point to this.  It seems to me you're trying to
>> build a system merely to do what VNC can do already!!
>> 
>> Of course, I could be wrong...
>
>You could be.

Or I could not be.

>Lot's of professionals run IBM's VM under VM for various reasons 
>including testing new releases.

That describes a *different* situation than what the original e-mail
described.  You describes virtualization on virtualization.  He
described running an effectively pointless operating system upon which
to run a Plex86 environment.

>Plex86 on IA32 would allow one to run a (possibly different) 
>distribution of Linux as a guest in a protected environment. I could 
>give you a Linux environment on my Linux machine with no possibility 
>that you could escape.

That's true.  That's a perfectly valid idea.  Why not run one distro as
the host OS, one (or more) as the guest?

Or, build a microkernel-based OS?  But why would you want to run an
effectively pointless OS, then run two distros under Plex86?  What's
the point of virtualizing both of them, when you only have to
virtualize one of them?

*That's* the point that I don't understand.  VM-under-VM gives you a
full-blown OS at the top. 
Full-Linux-under-Pontlessly-stripped-down-Linux doesn't give you
anything that Full-Linux-under-Full-Linux gives you.  *That* is my
point.

>If there be ways of improving its performance, many people will find 
>those advantageous.

The problem with performance and Plex86 is that Intel processors are
not completely virtualizable, as are mainframes.  Running a thin OS
doesn't really help you there.  If you want more than one OS accessing
the same piece of hardware (hard drives come to mind), then you will
have to battle against the IA32 architecture.  Running a thin OS
doesn't really help you there.

Tim Massey



Reply via email to