RE: [fw-wiz] Mainframes on the Net?

From: Paul D. Robertson (
Date: 11/15/02

From: "Paul D. Robertson" <>
To: <>
Date: Fri Nov 15 10:06:01 2002

On Fri, 15 Nov 2002 wrote:

> IIRC there were numerous ways to run Unices on IBM/3x0 and numerous Unices
> that ran there. Some were IBM-originated and some third-party ones.
> I've seen some of those long before posix stuff appeared in OpenMVS, IIRC it was
> late 80's. Don't know if those existed in Soviet Union were localized copies of
> something or written from scratch.

It's my recollection that the IBM-supplied ones were VM-based, not MVS
based, but that could just be a fuzzy recollection. One of the reasons
that I still like VM is that the hosting OS really just acts like a
monitor and virtualization realizer for the guests, providing very little
of its own services, and none of them (not memory, not disk, not
communications channels...) are provided without going through the
virtualization layer. Guests' programs just don't get to misbehave and
make SVC calls of the parent without some really gnarly and purposeful
compromise of the parent. Each user/daemon-like process having their own
virtual devices and virtual CPU/menory made things very well seperated.
Adding another OS to the virtual machine shouldn't require a change of the
host OS (indeed, we used to run OS/VS1 and VSE under VM quite reliably.)

> (nostalgic: i wonder how much old p/390 or russian ES-12xx cost these days?)

I never saw a P/390, but I alpha-tested the GA version of the P/370 for
both IBM and my employer at the time. It was something else to have those
huge bus and tag cables snaking out the pack of a PS/2 model 90 going to a
9 track tape drive :) The model 90 was the size of a PC server case, the
tape drives were about 5 feet high and 3' wide- not including the
controller, which also dwarfed the PC, though by not quite as much :)

The P/370 required a microchannel host and OS/2 (there was an AIX hosted
version on paper, not sure if it existed when I did my testing.) They
weren't that expensive back then, but the OS license might have been the
killer (well, that and the fact that they weren't all that fast- ~2MIPS.)
When I tested it, it wouldn't run MVS yet, but it would happily run VM
under VM, but only block devices (tape, 337x disks, 3725 comm.
controllers) would work- they were still working on 3380 drives.

Kind of interesting that we've gone from mainframes and Unix, to
mainframes running Unix, to Unix running mainframe emulation to mainframes
running Unix.

Paul D. Robertson "My statements in this message are personal opinions which may have no basis whatsoever in fact." Director of Risk Assessment TruSecure Corporation