Re: Kernel based virtual machine

[jtowler%soncom.com@localhost (John R. Towler)]

Jean-Yves Migeon <jeanyves.migeon%free.fr@localhost> writes:

> Le 14/10/12 21:42, Lukas Laukamp a écrit :
...
>
> IMHO porting the whole "turn the kernel into an hypervisor" in NetBSD
> will not really bring anything useful:
>
> - NetBSD already runs a dom0 Xen; ease of use does not depends on a
> choice of technology, more on a matter of tight integration of tools
> and what they can offer to you. See VMWare. They were talks to turn
> Xen into a loadable module, but I lost track of it.
>
> - KVM is fairly close to Linux kernel, and so are its
> internals. Unless you can dedicate man power to track and maintain
> these (like SmartOS), you have one more component that is bitroting on
> a daily basis. There is an attempt to do that though: BHyve.
>
> - from a system and security standpoint, I believe that the hypervisor
> should be the only privileged component on the system. The current
> "state of the art" adds the whole dom0 to the TCB, making this point
> moot. But you can design an OS where dom0 is kept to a minimum
> (Qubes).

I am not up on all the details and full knowledge of the design
decisions and major and minor chunks of system src code for this process
et al.  Some of that has to do with maintaining the known or perceived
differences in design from NetBSD as a BSD-4.4 system vs Linux and
Solaris as cited.  The author in this process moved from these issues to
the fully general question of OS redesign here involving NetBSD.

I had come across earlier, in finding out about Mach, and microkernels
and their issues, references to subsequent research institutions working
further with how this is/was involved with Unix design (say at the time
of the market prominence of the NeXT).  I recently checked to see what
was going on as a side interest.  The Univ. of Utah, CS dept. Flux OSKit
project etc. have moved from this research question to a project around
a kit of tools for (poss. visualizing) and analyzing the structure and
design of a Unix -type or other operating systems, their components, and
the services and etc. an OS provides.  I had thought I had seen this in
the pkgsrc tree as something interesting one night, so maybe this is an
obvious and  annoyingly redundant suggestion.  On rechecking today, I
did not find it via likely substring searches.

As an aid to how what is already known in the heads of experienced and
knowledgable people about NetBSD design, BSD-4.4 design, and where to
find and work with these at the right level immediately in the src tree,
including this discussion, the tools or ones like them of the Flux OSKit
might be an additional aid to research in the intended design direction
schema and testing its resulting functionality --- if this is not
already known, and people who know are not already involved with this.


<http://www.cs.utah.edu/flux/oskit/LICENSE>

<http://www.cs.utah.edu/flux/oskit/flux.html>

John R. Towler
jtowler%soncom.com@localhost