[vbox-dev] who is Cadmus Computer Systems?

Malcolm Scott Malcolm.Scott at cl.cam.ac.uk
Wed Aug 24 10:21:00 PDT 2011

Sorry to jump in very late here, but:

Frank Mehnert <Frank.Mehnert at ...> writes:

> On Tuesday 05 January 2010, Stéphane Charette wrote:
> > When I try to do a OUI lookup on the VirtualBox MAC addresses, I get
> > this strange entry:
> >
> > http://standards.ieee.org/regauth/oui/index.shtml
> >
> > 08-00-27   (hex)		CADMUS COMPUTER SYSTEMS
> > 080027     (base 16)		CADMUS COMPUTER SYSTEMS
> > 				600 SUFFOLK ST
> > 				LOWELL MA 08154
> >
> > I'd have expected maybe innotek GmbH, or Sun, or ...?  Now I'm curious
> > to know what is the story.
> innotek did never register an official MAC prefix. Registering
> a MAC prefix for VBox is on our todo list. Same for the PCI IDs.

Does anyone know the history behind the use of Cadmus's OUI?  I really hope it
is not the case that VirtualBox is "borrowing" another manufacturer's MAC
addresses without some agreement being in place; that would be very bad and
could have unpleasant consequences if VirtualBox ever shares a LAN with a real
Cadmus device.  Given the way in which VirtualBox uses the Cadmus OUI, by
effectively generating a random suffix, I can only see this use being valid if
Cadmus had never manufactured any device using this OUI and Cadmus's current
owners have agreed never to do so in the future.

If such an agreement is not in place, the best solution would be for VirtualBox
to obtain its own OUI (as VMware did), but that costs
(http://standards.ieee.org/faqs/regauth.html#14).  Failing that, the correct
alternative is to use locally-administered addresses (LAAs), i.e. any MAC
address with the 2nd LSB of the first octet set (XenServer does this).

Malcolm Scott
University of Cambridge Computer Lab

More information about the vbox-dev mailing list