A tour through the machineroom in october 2003

The machine room had been setup in the same place for almost three years, and had been remodeled a bit, with the arrival of new sets of shelves. Things are much more crowded than in the previous pictures now!

Standing in the doorway, this is what you can see: nothing. Almost nothing.
Because you're just in front of a dual set of shelves, over 6 feet tall (2 meters exactly), which go from the entrance to almost the other side of the room (really, they are only about 6 feet wide only).
Then, to the right, various large piece of hardware can be seen... almost nothing. Only the emerging part of the iceberg...

Don't enter the machineroom yet. Simply bend your head inside, and look to your right. This is what I call a small stack.
Why small? Well, count the machines. Count'em. Two to the right across the wall. Wait, three! The black thingie is a closed laptop. Five to the left. Wait, four! The white brick with blinking lights is only a mere UPS. So that makes seven machines, right? Oh, wait, one more to the left. The one with disk drive platters stacked on top of it.
So that makes eight machines. Not even a two-digit number. Of course it's small!

Remember, from the year 2000 pictures, the tower of Babel in a recess? Well, it's still there. No overall picture, there would not be enough details anyway...

Straight on the ground, three hp300 desktop machines (HP9000/4{00,25}t) share some room with a so-called shoebox Sun disk enclosure, which contains a huge 2GB SCSI disk.
First row, the black DSL modem is partly covered by dust, my bad. To its left, stand the MicroVAX which used to be this webserver, with an old iron (and heavy) HP 9000/720 machine on top of it (one of the first HP PA-RISC machines, successors to their HP9000/[34]00 line of m68k-based workstations). On top of the 720 itself lies an old SGC frame buffer suitable for this very machine... as well as for other HP machines.

Moving up, the second row welcomes a set of four Sun 3/60 machines, which were heavily involved back in the OpenBSD/sun3 days. The second one from the bottom is connected to the SCSI enclosure on the ground you saw in the previous picture. It used to run recent, uncommited code from my private Sun3 tree, which never got polished and commited.
One row above lies the strategic part of my home network, since the three Sparc cases to the left are my gateway, my fileserver and my main desktop machine, from bottom to top. There are also two hubs here, one 10baseT (RJ45) on the top left, and one 10base5 (AUI) on the bottom right. This is not a joke - AUI hubs did exist, and I was the lucky owner of one. This is way better than using 10base2 on non-10baseT-capable hardware, and saves me a lot of transceivers...
The next (and final) row sports one small (3U) VME card cage to the left, and two more Sparc machines as testbeds, to the right.

Oops. Looks like I was wrong, there are even more machines there. The upper (and really final) row welcomes a few more machines and unidentified books.
These machines are actually more HP PA-RISC workstations, seen from the rear side. The bottom one having much more horsepower than the others, I only use the upper two for occasional testing.

Time to look back to the left wall, which was almost blocking the entrance! Well, it's still a mess.
The left set of shelves has been set up to allow three large crates on the ground, which are two noisy and Watt-greedy 12U VME card cages, and a tower ROMP machine. Needless to say, the horizontal surface on top of these machines is not wasted, and at the time the pictures were taken, was filled with my set of Gravis Ultrasound soundcards, various uncommon mice, and random pieces of hardware, including a quarter-height 5"1/4 Micropolis hard drive...

Above the crates, from left to right: yet another unused 12U VME crate, used as storage for not-frequenly-used VME boards, one Sun JavaStation, and the monitor of the ROMP machine.

Looking above again, the next row is still crowded: to the left, an Alpha (DEC 3000/600), with an IBM RS/6000 model 320H on top of it (for a whooping 25MHz!), and a binder of various operating systems CD-ROM, as well as random hp300 memory sticks, and rubber gloves (at that time, I thought this would be a good idea to wear gloves while working on the innards of some systems). To the right, two small SCSI enclosures contain disks for the row of pmax machines above, with the tape drive of the ROMP machine on top of them; and in-between, more unprotected CD-ROM, as well as a few unsorted MCA cards.

One row up, the area is filled with four pmax (MIPS processor-based Digital machines, between the VAX and the Alpha lines), as well as various parts: hard drives, serial cables, more CD-ROMs, packed EEPROMs...
On the top of the shelves, boxes of parts or floppy disks; and to the left, original packaging for SunOS tapes and VMS 8" floppies.

Time to look at the right part of the left wall shelves now. The bottom is a mess, with a small UPS, an opened HP 9000/425t machine, with boards around, including an MVME162 to the left, and an HP 9000/743 with cables to the right.
On the first real row, keyboards cover various external SCSI enclosures; the right keyboard is really an half-dead (thus unreliable) Amiga 1200 with a Blizzard 1230 accelerator, which I am not using anymore... (and to add to the sad story, this machine got lost while relocating a few years later).

Looking above, the coast is clear: this time, the row is only populated with three monitors, each with a different connector, for different machines around.
Err, clear? That is, if you don't look at the front of the monitors. Which is filled with random HIL devices (the nine-knob device, which is really, really, really, from a software point of view, three distinct three-axis mice), more HP 9000/743 cables, MVME memory boards, etc.
Oh, and before someone asks, the black motherboard suspended from the shelves, of which you can see the right half in this picture, is an Apollo DN3500 motherboard. This is an impressive piece of engineering, based upon a Motorola 68030-68882 couple, yet a 7-slot ISA bus, and a specific connector for an half-ISA, half-proprietary floating point accelerator. From Weitek of course, who else would manufacture FPA back then? Oh, the year? 1987. When Apollo could have been rocking the Unix world, had their operating system been a real Unix...
Back in the mood for the visit, the next row is not much tall, and only contains an Acorn RiscPC 610 to the left (which revolving front door does not stay in the upper position anymore), and an Apple Macintosh Quadra 650 - one of the best m68k-based Macintoshes - to the right. Yes, that's true! After telling people for years I would never have any Apple hardware at home, I eventually bought this machine (for an insanely low price, of course). I wish I had bought a Quadra 840AV, which runs at 40MHz instead of 33MHz, but its case is much larger, so the Quadra 650 was a good compromise, given my reluctance to play on a m68k-based Macintosh machine. After all, despite A/UX, and the various free operating systems ports to this hardware, most Macintoshes still stink of an early 1980s design, which was pretty good for the time and the goal of a diskless, all-in-one, desktop computer, but are tragically bad (some days I say wrong) from a Unix machine point of view...

Above again, the remaining space does not look really filled - to the left, two half-open, half-dead, peecee boxens are filled with various hardware in unknown condition, and to the right, the BeBox looks like it tried to flee to a better place...
The left box is (or was) actually a fully-functionnal peecee, which I used for various driver tinkering, especially ISA drivers. The other one, closer to the center of the shelves, used to be, a very long time ago, an OS/2 file server for my home network, providing TCP/IP and NFS services since the OS/2 2.10 days. Those were the days! I have written so many lines of code on it...
In-between the OS/2 peecee and the BeBox, the white thing is an old Sun SCSI enclosure, featuring two full-height 5"1/4 bays. This one is filled with an SCSI disk and a QIC-150 tape drive, and I use it as a recovery media for Sparc machines with dying drives...

Finally, the upper part of the right shelves. No surprise, boxens and accessories again. Did you notice the parts above the left peecee of the previous row? Apart from an ISA network card, there is a DN3500 motherboard again! With the red power supply connector...

The RS/6000 550 is still here. And I am still putting too many things on top of the monitors... What's here? Disk drives. MVME boards. Various connectors. Adhesive tape. And whatnot!

The brown Sun VME crates are still there, too. As well as the grey ones. Here, across the right wall, you can see the back of three of the now four brown crates (here, right to left, 3/260, SMD disks, 4/260). The SMD drive cabinet is open, you can see the back plate with two fans and cables on the left, and the drive themselves are barely visible inside the cabinet. In the back, the 670 crate is almost at the same location. But the 330 crate has been moved to the other side of the desktop.

Before moving back to my desktop, here is a look at the various machines and accessories on top of the brown VME crates. While the Sun Ultra 5 is almost invisible, the small black UPS on top of it, as well as the Sun Ultra 1/Creator (1E) are easily recognizable.
But that's nothing compared to the few MVME board assemblies lying on top of the U1E, including various ready-to-run multiprocessor MVME188A assemblies. Note, to the upper right corner of the U1E, a stack of Motorola HYPERmodules for the MVME188 series.

Here is a close-up on the VME rack in the tower of Babel. Back then, there were only two boards in it. An MVME177 (Motorola 68060-based) at the bottom, with two stacked 32MB memory boards (hence two slots), in the bottom. And an MVME187 (Motorola 88100-based) at the top. Not having the drive racks to suspend the hard drives from the top of the enclosure, I simply put the drive on top of antistatic foam.
Nowadays, I have replaced the two 32MB memory boards with a single 64MB board, and put a Momenco Leopard-V UltraSparc board in the middle (courtesy of Jason L. Wright). And one more internal drive. And one more external drive.

Remember the blue Alpha three years ago? Well, I still use it. Here is a close-up of its front panel. See the glued i486DX CPU? Well, while this machine was travelling a lot with me, from '98 to '01, I kept encountering people asking me whether this was the processor of the machine... I was evil enough not to contradict some of them...
Too bad the black paint, which is supposed to be a bit shiny, is having a bad time those days...

My everyday use of the machineroom mainly involves two machines. One of these is an SGI Indigo 2, now running at 250 MHz (woo hoo!), connected to the granite keyboard and monitor to the left of the desktop. The monitor is currently displaying the Octahedra OpenGL screensaver.
Don't pretend not to notice the red Beastie puppet !
And while you're here, to the far left, the black laptop is a Tadople SparcBook 3GX computer, courtesy of Todd Miller, on which I did most of my Sparc development work from late 2003 to late 2005.

The pine cone above the middle monitor is almost the canonical decoration here. I love huge pine cones, and I am glad to own a bunch of them.
Of course, pine cones have the lowest priority of all useless gizmos around, which explains why so many monitors have strange things on top of them, such as the right monitor here. Can you catch everything on top of it? 8 metal rings, from a Fujitsu hard drive. One dead VAXstation graphics board. One dead UltraSparc I/167 processor. One white rubber advertising computer-shape thingie I got from an o[dl]d job. And even more...

What a high density of cables! Here are about 20 10baseT network cables, a couple of serial cables, small SCSI cables, and even more. But no knots!

Remember the VMS 1.6 floppy disks, too? They are still here. And this is a better picture of them.
This is a part of computing history.