1 Feb
2013
1 Feb
'13
9:42 p.m.
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
A fellow contacted me with this problem:
I have a couple of Kurzweil synthesizers (/samplers/sequencers):
K2000RS and K2600X . They have integrated floppy and SCSI HDDs
for storing configuration, composition and audio sample data.
They're actually 68000 CPUs with a Kurzweil OS,
managing IO among peripherals
(keys, SCSI, display LCD, MIDI, etc) and several proprietary DSPs.
But there's no serial/network/etc IO.
I'm interested in trying to use their SCSI
to interface them with a (Linux) PC somehow.
Any ideas? Like maybe some webpages
describing projects that have networked older
(preferably 68xxx) CPUs over their shared SCSI bus.
My understanding is that a Kurzweil and a Mac were sometimes configured
each as a SCSI device on the same SCSI bus as the HDD,
and either host could alternate accessing the HDD
once the other host was no longer issuing
potentially conflicting SCSI commands to the HDD.
I don't see how just read commands could conflict,
but perhaps the HDD wasn't able to multiplex commands
targeting different hosts in any overlap.
But evidently people were transferring samples
between Kurzweil and Mac this way.
[I replied privately about the problem
of multiple machines sharing a drive even if one is read-only]
any clues or hints?
thanks
-- jeff jonas
2 Feb
2 Feb
12:54 p.m.
I'm interested in trying to use their SCSI
to interface them with a (Linux) PC somehow.
Any ideas? Like maybe some webpages
describing projects that have networked older
(preferably 68xxx) CPUs over their shared SCSI bus.
As far as I remember, you can have multiple hosts on line, if they are
not supplying current to the active terminators, and uses different SCSI
IDs...
9:52 p.m.
Alexandre Souza - Listas wrote:
This would be interesting to some folks I know who play with test gear
that runs on OS9, has a SCSI port on the rear, but that the software
only considers a tape drive hooked up to that port as valid (Fluke 9100
series). Anyone here work with OS9?
John :-#)#
I'm
interested in trying to use their SCSI
to interface them with a (Linux) PC somehow.
Any ideas? Like maybe some webpages
describing projects that have networked older
(preferably 68xxx) CPUs over their shared SCSI bus.
As far as I remember, you can have multiple hosts on line, if they
are not supplying current to the active terminators, and uses
different SCSI IDs...
1:11 p.m.
On 02/01/2013 09:42 PM, Jeff Jonas wrote:
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
One used to be able to do it with Suns, in the Good Ol' Days. I've
not done it myself but have seen it done. It was really quite neat.
A fellow contacted me with this problem:
I have a couple of Kurzweil synthesizers (/samplers/sequencers):
K2000RS and K2600X . They have integrated floppy and SCSI HDDs
for storing configuration, composition and audio sample data.
They're actually 68000 CPUs with a Kurzweil OS,
managing IO among peripherals
(keys, SCSI, display LCD, MIDI, etc) and several proprietary DSPs.
But there's no serial/network/etc IO.
I'm interested in trying to use their SCSI
to interface them with a (Linux) PC somehow.
Any ideas? Like maybe some webpages
describing projects that have networked older
(preferably 68xxx) CPUs over their shared SCSI bus.
My understanding is that a Kurzweil and a Mac were sometimes configured
each as a SCSI device on the same SCSI bus as the HDD,
and either host could alternate accessing the HDD
once the other host was no longer issuing
potentially conflicting SCSI commands to the HDD.
I don't see how just read commands could conflict,
but perhaps the HDD wasn't able to multiplex commands
targeting different hosts in any overlap.
But evidently people were transferring samples
between Kurzweil and Mac this way.
Read commands could conflict in the simple case of a host assuming it
has exclusive control of the bus. I don't believe any reasonable SCSI
host adapter chips would behave that way, but some early SCSI host
adapters were implemented using discrete logic or with a simple
interface that needs lots of OS handholding (like the NCR 5380) and
high-level interactions are done in the driver.
Unless you want to dig deeply into the Kurzweil firmware, I don't
really see this working out, though. This is assuming you don't have
any documentation on how the Mac interacts with the synthesizer or the
on-disk formats in use.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
8:11 p.m.
On 02/02/2013 18:11, Dave McGuire wrote:
On 02/01/2013 09:42 PM, Jeff Jonas wrote:
A company called Lingenuity did the same with a product called SCSI
Share, for Acorn computers, around 1989 or 1990. It allowed six
computers to share a SCSI drive, which appeared as two partitions, one
read-only for software and one read-write for shared files. Of course
distances were limited by the SCSI bus, which meant you needed decent
cables and all the devices more or less on the same tabletop. But I
have seen it working on more than one occasion.
--
Pete Peter Turnbull
Network Manager
University of York
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
One used to be able to do it with Suns, in the Good Ol' Days. I've
not done it myself but have seen it done. It was really quite neat. 
3 Feb
3 Feb
4:40 p.m.
On Feb 2, 2013, at 5:11 PM, Pete Turnbull wrote:
On 02/02/2013 18:11, Dave McGuire wrote:
Did this while I was at IBM. We typically had demos of 2-3 systems accessing a
number of shared disks.
It's called multi-initiator mode. Though it was supposed to work (as per spec), not
all devices were great at having multiple initiators (hosts) on the same bus.
Slightly different, Adaptec did produce a device called a "Nodem" which was a
SCSI device that connected to ethernet. Was pretty cool in concept. However, I spent a
lot of time working with Adaptec and it has some serious bugs that they could never quite
work out. The biggest was they had a minimum interpacket time in order to receive an
ethernet packet. Unfortunately, the Nodem's interpacket time was longer than the
minimum interpacket time on ethernet. This resulted in some cases of missed packets.
:-(
TTFN - Guy
On 02/01/2013 09:42 PM, Jeff Jonas wrote:
A company called Lingenuity did the same with a product called SCSI
Share, for Acorn computers, around 1989 or 1990. It allowed six
computers to share a SCSI drive, which appeared as two partitions, one
read-only for software and one read-write for shared files. Of course
distances were limited by the SCSI bus, which meant you needed decent
cables and all the devices more or less on the same tabletop. But I
have seen it working on more than one occasion.
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
One used to be able to do it with Suns, in the Good Ol' Days. I've
not done it myself but have seen it done. It was really quite neat. 
4 Feb
4 Feb
1:01 p.m.
On Feb 3, 2013 9:49 PM, "Guy Sotomayor" <ggs at shiresoft.com> wrote:
On Feb 2, 2013, at 5:11 PM, Pete Turnbull wrote:
accessing a number of shared disks.
On 02/02/2013 18:11, Dave McGuire wrote:
Did this while I was at IBM. We typically had demos of 2-3 systems
On 02/01/2013 09:42 PM, Jeff Jonas wrote:
A company called Lingenuity did the same with a product called SCSI
Share, for Acorn computers, around 1989 or 1990. It allowed six
computers to share a SCSI drive, which appeared as two partitions, one
read-only for software and one read-write for shared files. Of course
distances were limited by the SCSI bus, which meant you needed decent
cables and all the devices more or less on the same tabletop. But I
have seen it working on more than one occasion.
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
One used to be able to do it with Suns, in the Good Ol' Days. I've
not done it myself but have seen it done. It was really quite neat.
It's called multi-initiator mode. Though it was supposed to work (as per
spec), not all devices were great at having multiple initiators (hosts) on
the same bus.
Slightly different, Adaptec did produce a device called a "Nodem" which
was a SCSI device that connected to ethernet. Was pretty cool in concept.
However, I spent a lot of time working with Adaptec and it has some
serious bugs that they could never quite work out. The biggest was they
had a minimum interpacket time in order to receive an ethernet packet.
Unfortunately, the Nodem's interpacket time was longer than the minimum
interpacket time on ethernet. This resulted in some cases of missed
packets. :-(
TTFN - Guy
They had scsi<->ethernet adapters for Macs that didn't have expansion slots
I think Asante made them.
1:20 p.m.
On Mon, Feb 4, 2013 at 10:01 AM, John Many Jars
<john at yoyodyne-propulsion.net> wrote:
They had scsi<->ethernet adapters for Macs that didn't have expansion slots
I think Asante made them.
One of those just showed up on eBay this morning. Not something I
follow and I have no idea what a reasonable value is for one of these.
But it's "Rare!"
Asante ENSC SCSI To Ethernet Adapter Rare!
http://www.ebay.com/itm/290857727985
$45, BIN $75
1:38 p.m.
On Feb 4, 2013, at 1:20 PM, Glen Slick <glen.slick at gmail.com> wrote:
On Mon, Feb 4, 2013 at 10:01 AM, John Many Jars
<john at yoyodyne-propulsion.net> wrote:
I found it handy for a Mac SE, back in the day when an SE Ethernet
card cost an arm and a leg and one of these was lying around spare
for some reason. These days, unless I needed something else in the
SE expansion slot, I'd probably just go for that, because they were
going for about $30 last I looked (which was, to be fair, about 4
years ago).
If you want to attach a Mac Plus to any kind of modern network
without a seriously quirky LocalTalk bridge, it's pretty much your
only bet. Same applies to non-color Classics, which didn't have
expansion slots (the Color Classic line had LC-PDS slots). I can't
think of any other non-expandable Macs that didn't have Ethernet
built in. I suppose that's why it's "R at RE!!!!"; its market was
extremely limited, especially since there probably weren't loads of
Classic users trying to get onto Ethernet at the time.
- Dave
They had scsi<->ethernet adapters for Macs that didn't have expansion slots
I think Asante made them.
One of those just showed up on eBay this morning. Not something I
follow and I have no idea what a reasonable value is for one of these.
But it's "Rare!" 
1:40 p.m.
On Mon, Feb 4, 2013 at 1:20 PM, Glen Slick <glen.slick at gmail.com> wrote:
Asante ENSC SCSI To Ethernet Adapter Rare!
http://www.ebay.com/itm/290857727985
$45, BIN $75
Anyone know where to get installer disks for those? I have something
like it (similar, but in a different box, so could be a different
model) and no software.
-ethan
1:54 p.m.
On Feb 4, 2013, at 1:40 PM, Ethan Dicks <ethan.dicks at gmail.com> wrote:
On Mon, Feb 4, 2013 at 1:20 PM, Glen Slick
<glen.slick at gmail.com> wrote:
I can look to see if I have it in my pile of SE disks, or at least
the extension (which might still be on the machine). I lost track
of the device a long time ago, back when it seemed less important.
- Dave
Asante ENSC SCSI To Ethernet Adapter Rare!
http://www.ebay.com/itm/290857727985
$45, BIN $75
Anyone know where to get installer disks for those? I have something
like it (similar, but in a different box, so could be a different
model) and no software.
1:56 p.m.
On Feb 4, 2013, at 1:40 PM, Ethan Dicks <ethan.dicks at gmail.com> wrote:
On Mon, Feb 4, 2013 at 1:20 PM, Glen Slick
<glen.slick at gmail.com> wrote:
Though in theory, this should work as well (there's a disk image at
the end):
http://www.jagshouse.com/PlusToiMac.html
- Dave
Asante ENSC SCSI To Ethernet Adapter Rare!
http://www.ebay.com/itm/290857727985
$45, BIN $75
Anyone know where to get installer disks for those? I have something
like it (similar, but in a different box, so could be a different
model) and no software.
2:18 p.m.
On Feb 4, 2013, at 1:56 PM, David Riley <fraveydank at gmail.com> wrote:
On Feb 4, 2013, at 1:40 PM, Ethan Dicks
<ethan.dicks at gmail.com> wrote:
And it appears that Asante still has downloads up, which is actually
astonishing at this point:
http://www.asante.com/support/legacy/Bridges/EN_SC.asp
- Dave
On Mon, Feb 4, 2013 at 1:20 PM, Glen Slick
<glen.slick at gmail.com> wrote:
Though in theory, this should work as well (there's a disk image at
the end):
http://www.jagshouse.com/PlusToiMac.html Asante ENSC SCSI To Ethernet Adapter Rare!
http://www.ebay.com/itm/290857727985
$45, BIN $75
Anyone know where to get installer disks for those? I have something
like it (similar, but in a different box, so could be a different
model) and no software.
2:43 p.m.
On Mon, Feb 4, 2013 at 2:18 PM, David Riley <fraveydank at gmail.com> wrote:
And it appears that Asante still has downloads up,
which is actually
astonishing at this point:
http://www.asante.com/support/legacy/Bridges/EN_SC.asp
Thank you! I'll have to dig out my Mac accessory box and see what model I have.
-ethan
6 Feb
6 Feb
3:33 a.m.
On Mon, Feb 4, 2013 at 2:43 PM, Ethan Dicks <ethan.dicks at gmail.com> wrote:
On Mon, Feb 4, 2013 at 2:18 PM, David Riley
<fraveydank at gmail.com> wrote:
It looks like I do have the model in the eBay listing posted earlier,
the old boxy
EN/SC. Internally, there's a Natl Semi DP8390DN for Ethernet, and for the
media interfaces, a DP8391AN and DP8392CN. I can't find anything that
resembles a microprocessor and it looks like the SCSI interface is implemented
on a trio of PALs. There happen to be some counters (74LS393) between the
PALs and the 27C256 EPROM, so perhaps there's just a simple state machine
that pushes and pulls magically-formatted SCSI packets and all the rest of
the smarts are in the driver.
I'm missing the power adapters and the boxes aren't marked. Another model
uses 12VAC wall warts. This PCB has a big fat 2A block rectifier, so perhaps
it's the same. There's no linear regulator like a 7805 on there, but there's
a
Motorola MC34163P and a big, fat coil, so it appears it has its own switching
regulator built into the board.
I think from a quick read of the Asante docs that it will work with
System 6.0.8 or newer, meaning at least 6.something. IIRC, Apple
made significant changes in networking over System 7.
Thanks again for the pointers. I think I have enough to put these to use
on an old Mac Plus or Mac SE.
-ethan
And it appears that Asante still has downloads
up, which is actually
astonishing at this point:
http://www.asante.com/support/legacy/Bridges/EN_SC.asp
Thank you! I'll have to dig out my Mac accessory box and see what model I have.
4 Feb
4 Feb
12:47 p.m.
At 12:11 PM 2/2/2013, Dave McGuire wrote:
On 02/01/2013 09:42 PM, Jeff Jonas wrote:
There was an Amiga product like that, once upon a time. It was
short-lived. It was made by C Ltd.
- John
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
One used to be able to do it with Suns, in the Good Ol' Days. I've
not done it myself but have seen it done. It was really quite neat. 
2 Feb
2 Feb
2:25 p.m.
Jeff Jonas [jeffj at panix.com] wrote:
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
I don't remember the details but I'm fairly sure you could run
OpenVMS clusters over SCSI at some point. I don't recall
whether this was "just experimental" or whether it was
actually supported in some configurations.
Antonio
arcarlini at iee.org
3 Feb
3 Feb
7:37 a.m.
On 03/02/2013, at 6:25 AM, arcarlini at iee.org wrote:
Jeff Jonas [jeffj at panix.com] wrote:
Fully supported on OpenVMS/Alpha with the appropriate adapters - I certainly set up more
than one such cluster. IIRC it's recently been qualified for Integrity as well.
Being OpenVMS the disks are shared active/active between the two nodes, all you had to do
was set the SCSI ids of the two controllers to a different value (suggested was 7 for one,
6 for the other).
Huw Davies | e-mail: Huw.Davies at kerberos.davies.net.au
Melbourne | "If soccer was meant to be played in the
Australia | air, the sky would be painted green"
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
I don't remember the details but I'm fairly sure you could run
OpenVMS clusters over SCSI at some point. I don't recall
whether this was "just experimental" or whether it was
actually supported in some configurations. 
2 Feb
2 Feb
4:32 p.m.
On 02/01/2013 09:42 PM, Jeff Jonas wrote:
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
No rare was the need for multiple loosely coupled Ampros.
First the interface used by Ampro was a loose SCSI (pre SCSI spec!)
And it could be used as a generalized bidirectional parallel bus
or it could so SASI (sasi was the pre SCSI) and with the right code
it could do slow SCSI (programmed IO). I used it to connect to a
3.5" fujitsu 45MB SCSI drive. Worked well lots of code needed
the base drivers only did a very limited subset for use with the
Adaptec SCSI to MFM bridge board or the ST225N.
So there is no resemblence there to SCSI as we currently know it.
A fellow contacted me with this problem:
I have a couple of Kurzweil synthesizers (/samplers/sequencers):
K2000RS and K2600X . They have integrated floppy and SCSI HDDs
for storing configuration, composition and audio sample data.
They're actually 68000 CPUs with a Kurzweil OS,
managing IO among peripherals
(keys, SCSI, display LCD, MIDI, etc) and several proprietary DSPs.
But there's no serial/network/etc IO.
I'm interested in trying to use their SCSI
to interface them with a (Linux) PC somehow.
Any ideas? Like maybe some webpages
describing projects that have networked older
(preferably 68xxx) CPUs over their shared SCSI bus.
My understanding is that a Kurzweil and a Mac were sometimes configured
each as a SCSI device on the same SCSI bus as the HDD,
and either host could alternate accessing the HDD
once the other host was no longer issuing
potentially conflicting SCSI commands to the HDD.
I don't see how just read commands could conflict,
but perhaps the HDD wasn't able to multiplex commands
targeting different hosts in any overlap.
But evidently people were transferring samples
between Kurzweil and Mac this way.
[I replied privately about the problem
of multiple machines sharing a drive even if one is read-only]
any clues or hints?
Serious memory test here...
All SCSI devices have an address including the host and the maximum was 8
total (0-7). Also the bus assumed only one controller so all other
units were
slaves that could do no more than assert ATN
Early (that era) SCSI disks did not know who issued command as they were
slaves. Further they were FIFO serial command structure with minimal if
any command silo and the assumptuion that the command executed
produced results that went back to the only host controller.
Are you sure the bus was SCSI and not GPIB? GPIB was a multi talker,
multi controller, multi-listener bus that allowed for up to 31 devices
and was also a parallel 8bit bus.
Allison
thanks
-- jeff jonas
4:56 p.m.
later scsi technologies allowed double the devices.if you google scsi networking you get
this as a hit,a project called ipoverscsilooks reasonable
4:59 p.m.
any clues or hints?
Serious memory test here...
All SCSI devices have an address including the host and the maximum was 8
total (0-7). Also the bus assumed only one controller so all other
units were
slaves that could do no more than assert ATN
I am also pretty sure you could share disk between Ataris on the same
BUS. These days with reserve/release we do it all the time...
Early (that era) SCSI disks did not know who issued
command as they were
slaves. Further they were FIFO serial command structure with minimal if
any command silo and the assumptuion that the command executed
produced results that went back to the only host controller.
Are you sure the bus was SCSI and not GPIB? GPIB was a multi talker,
multi controller, multi-listener bus that allowed for up to 31 devices
and was also a parallel 8bit bus.
Allison
thanks
-- jeff jonas
4 Feb
4 Feb
4:46 p.m.
From: Jeff Jonas
Sent: Friday, February 01, 2013 6:43 PM
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
At XKL, I once mentioned the use of CI-based disks on TOPS-20 (major
feature of v6.0) to one of the SCSI card engineers for the Toad-1.
He was intrigued by the idea, read the docs and the TOPS-20 source
code for the feature, and implemented MSCP on the FASTWIDE SCSI bus.
It never made it past the experimental stage (unfortunately, in my
opinion).
Computer Interconnect (CI) was a 50Mbit ethernet[1] designed for the
Jupiter (follow-on to the DEC-20, to be called the DECSYSTEM-4050),
and later retrofitted to the VAX, so I would think that our little
experiment qualifies as an edge case. :-)
[1] Lower case is intentional.
Rich Alderson
Vintage Computing Sr. Systems Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
6:24 p.m.
On 2/4/2013 1:46 PM, Rich Alderson wrote:
From: Jeff Jonas
Sent: Friday, February 01, 2013 6:43 PM
CMD now Silicon Imaging, who have made a lot off of a PCI to IDE chip
had a CI disk subsystem which hooked to Vax systems. The chip they used
escapes me, but it was a SCSI chip of some sort that was persuaded to
parse the info coming from the VAX. I believe they dispatched action in
the controller within a couple of their clocks from having the complete
command received from the VAX. The DEC hardware didn't stand a chance
performance wise, as it had to read the command an chew on it for a
while in some processor before anything happened. If the Cache in the
CMD subsystem had the data it was turned around from the buffer starting
practically a soon as the command byte went out.
Jim
Long ago when SCSI was young
and just an 8 bit parallel bus,
Ampro littleboards advocated using the SCSI bus
for peer to peer communications, not just master/slave.
Did anyone else do that?
At XKL, I once mentioned the use of CI-based disks on
TOPS-20 (major
feature of v6.0) to one of the SCSI card engineers for the Toad-1.
He was intrigued by the idea, read the docs and the TOPS-20 source
code for the feature, and implemented MSCP on the FASTWIDE SCSI bus.
It never made it past the experimental stage (unfortunately, in my
opinion).
Computer Interconnect (CI) was a 50Mbit ethernet[1] designed for the
Jupiter (follow-on to the DEC-20, to be called the DECSYSTEM-4050),
and later retrofitted to the VAX, so I would think that our little
experiment qualifies as an edge case. :-)
[1] Lower case is intentional.
Rich Alderson
Vintage Computing Sr. Systems Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
6 Feb
6 Feb
7:10 a.m.
FWIW,
Digiboard here in MN, built a line of intelligent multi-port serial
adapters which connected to SCSI. The OEM'd them for SGI.
Chris
--
Chris Elmquist
11:57 a.m.
On Wednesday (02/06/2013 at 06:10AM -0600), Chris Elmquist wrote:
FWIW,
Digiboard here in MN, built a line of intelligent multi-port serial
adapters which connected to SCSI. The OEM'd them for SGI.
Just closing the loop on this, after checking with my drinking buddies
(aka, former Digi cohorts).
Turns out the devices were originally built by Central Data. Digi bought
Central Data and kept the line alive for a while. SGI had already
integrated them into IRIX and although they were called "SCSI Terminal
Server", they really presented a tty device interface to the OS and
application software. You ended up with a tty device per port in the
/dev hierarchy.
see eg,
http://techpubs.sgi.com/library/tpl/cgi-bin/getdoc.cgi?coll=0650&db=man…
Chris
--
Chris Elmquist
5:04 a.m.
Rich Alderson wrote:
Computer Interconnect (CI) was a 50Mbit ethernet[1]
70 Mbps.
Although it clearly was inspired by Ethernet, it is dissimilar enough
that I'm not sure I'd even call it a (lower-case "l") ethernet.
designed for the
Jupiter (follow-on to the DEC-20, to be called the DECSYSTEM-4050),
and later retrofitted to the VAX,
Perhaps you have more information about CI history than I do, but the
documents I've seen suggest that it was intended for everything from the
KL and VAX-11/780 up. The case could be made that it was "retrofitted"
to the VAX-11/750. The original CI hardware, including the PLI
interface, was clearly designed to be suitable for machines with memory
bandwidth as low as that of the VAX-11/780.
Eric
3:12 p.m.
From: Eric Smith
Sent: Wednesday, February 06, 2013 2:04 AM
Rich Alderson wrote:
> Computer Interconnect (CI) was a 50Mbit
ethernet[1]
70 Mbps.
Not originally. That was the HSC70 upgrade, which was VAX only.
Although it clearly was inspired by Ethernet, it is
dissimilar enough
that I'm not sure I'd even call it a (lower-case "l") ethernet.
It's similar enough that I would.
> designed for the
> Jupiter (follow-on to the DEC-20, to be called the DECSYSTEM-4050),
> and later retrofitted to the VAX,
Perhaps you have more information about CI history
than I do, but the
documents I've seen suggest that it was intended for everything from the
KL and VAX-11/780 up. The case could be made that it was "retrofitted"
to the VAX-11/750. The original CI hardware, including the PLI
interface, was clearly designed to be suitable for machines with memory
bandwidth as low as that of the VAX-11/780.
When I first heard about the CI, it was in a Jupiter dog-and-pony that DEC
did for us at UChicago. CI was touted as new for the Jupiter, and there
was no indication that it would be retrofitted to older 36-bit systems.
That was in 1982.
When TOPS-20 v6 was field-tested at Stanford, the Digital muffins were all
excited because the CI had been made available for the KL processors in
spite of the 1983 cancellation of Jupiter.
That was in 1986.
I think that any documentation from Digital which suggests that the CI was
intended for anything other than the Jupiter is revisionism in extremis.
Just my observations, of course. YMMV. OIMACTTA. And all that.
Rich Alderson
Vintage Computing Sr. Systems Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
8 Feb
8 Feb
10:50 p.m.
I wrote about CI, that it's rate was:
I'll agree that CI is slightly more similar to Ethernet than to Token
RIng. Both CI and Ethernet use Manchester encoding and a 32-bit CRC as
a FCS. However, CI does not use CSMA/CD arbitration, which is one of
the defining characteristics of Ethernet (aside from the newer
point-to-point Ethernet links, which have no arbitration). CI has its
own arbitration scheme based on node numbers, and does not use
exponential backoff. Not to mention that CI doesn't support a bus
topology, only a star. And a CI network is limited to a 90m radius,
while Ethernet was designed for a maximum station separation of 2.5Km.
So please remind me again just exactly what is so similar between CI and
Ethernet?
70 Mbps.
Rich Alderson wrote:
Not originally. That was the HSC70 upgrade, which was
VAX only.
The HSC70 upgraded several things, but not the raw data rate, which was
70 Mbps from day one. See, for example, the HSC50/70 Hardware Technical
Manual, EK-HS571-TM-001. The line rate is determined by the "Port Link
Module" (L0100). This same module is used in the HSC50, HSC70, and
CI20. The module is hardwired for 70 Mbps operation.
As far as I've been able to determine, the main performance boost of CI
network traffic with the HSC70 was that it supported longer packets,
reducing overhead. However, in most cases it was probably more
significant that due to the higher processor performance of the HSC70
vs. HSC50, it could handle more disk I/O operations per second.
It's definitely a shame that the HSC70 didn't support 576-byte sector
mode for use with 36-bit systems. This resulted from a deliberate
choice to drop 576-byte mode from the HSC70 software, as there is no
such limitation in the HSC70 hardware.
Although it
clearly was inspired by Ethernet, it is dissimilar enough
that I'm not sure I'd even call it a (lower-case "l") ethernet.
It's similar enough that I would. I think that any documentation from Digital which
suggests that the CI was
intended for anything other than the Jupiter is revisionism in extremis.
So the dates on the schematics in the Field Maintenance Print Set for
the CI20 are revisionist history? Fascinating.
Also the DEC internal memo "Clusters Program Analysis" memo by Tony
Sukiennik, dated June 18, 1982, states on page 13:
The CI is a high speed (70 megabits/second), multdropped, short
distance (90 meter radius) interconnect designed to pass data
and control infomration among intelligent computers.
The computers currently supporting the CI are as follows:
1. VAX-11/780
2. VAX-11/782
3. VAX-11/750
4. 2060
5. 2080 (JUPITER)
6. VENUS
7. HSC-50 (I/O server)
[Note that the presence of the VAX-11/750 on the list contradicts my
earlier suggestion that the CI750 might have been an afterthought.]
If it had really been the case that the CI20 wasn't planned until after
the Jupiter cancellation, there's no way that engineering could have
gotten it ready between then and when it first shipped. Just the
engineering of the three CI20 modules that are installed in the KL's
RH20 backplane (M3001, M3002, and M3003) appears to have taken at least
two years.
Eric
4187
days inactive
4194
days old
27 comments
20 participants
participants (20)
-
ajp166＠verizon.net -
arcarlini＠iee.org -
chrise＠pobox.com -
dave.g4ugm＠gmail.com -
dgahling＠hotmail.com -
eric＠brouhaha.com -
ethan.dicks＠gmail.com -
fraveydank＠gmail.com -
ggs＠shiresoft.com -
glen.slick＠gmail.com -
huw.davies＠mail.vsm.com.au -
jeffj＠panix.com -
jfoust＠threedee.com -
john＠yoyodyne-propulsion.net -
jwsmail＠jwsss.com -
mcguire＠neurotica.com -
pete＠dunnington.plus.com -
pinball＠telus.net -
pu1bzz.listas＠gmail.com -
RichA＠vulcan.com