On Fri, 1 Dec 2000, Al B. Snell wrote:

> This is the coolest I've seen yet:
> 
> http://www.general-micro-systems.com/v56.html

That *is* a nice board. It's a bit too different to the motorola mvme68k
boards for NetBSD/mvme68k to support it. However, if someone wants to send
me one, I could get a new port: "NetBSD/gmsvme68k" running on it within a
month or two... ;-)  (That's the theory anyway; in reality, I simply can't
spare the time now or for the foreseeable future).

> What's the deal with stuff wired to the P2 socket? How do you get the SCSI
> bus and the serial ports on that thing? Is that what a "transition
> board" is for?

Looks like it uses the same P2 pinout as Motorola's MVME712 rear
transition module. That's a board which connects to the back of the VMEbus
P2 connector and provides a panel with scsi, aui, printer and serial
ports.

> It'd be fairly neat for me to have a load of CPU cards on a backplane
> communicating TCP/IP over said backplane... that'd make a great
> extensible falloverable cluster.

Indeed, but VME devices usually command a premium price. I suspect the
"V56" would be in the $4000-$6000 range if GMS are competing with
motorola's offerings. Still, 

> Just imagine: as an Internet appliance (as made by www.cobalt.com),
> something where you can add "disk cards" (CPU, RAM, hard disk, serving NFS
> over the backplane), "CPU cards" (small disk for OS+swap) and "network
> cards" (routers) at will... mix and match... Mmmmm....

Yup, that's exactly what VME is designed/used for. There is already a
standard for IP over the VMEbus backplane, and implementing it for NetBSD
is somewhere in my work queue, though not near the top. 8-)

Cheers, Steve