Connecting multiple computers to one chain of SIO devices

[Timbit]

Is it possible to have two or more Atari 8-bit computers share a chain of SIO devices, either simultaneously or in turn? I'm thinking of something like the Handic VIC-Switch for the Commodore 8-bit computers.

I also asked this elsewhere and was pointed to the "Automatic Two Computer Interface" mentioned in the page linked below. It only supports two computers but, based on how it works, I think it could be expanded to support more. You will have to page down a few times to find it: http://www.mathyvannisselroy.nl/special%20stuff.htm

 

At the end of that section, he mentions the A2RI. I'm not sure whether this is the same device but it seems to have the same functionality. This is in German: http://www.mathyvannisselroy.nl/2r_bauan.PDF

 

Does anyone know whether anyone is selling such a device, or know of other options?

 

Thanks.

Edited March 25, 2015 by Timbit

[El Destructo]

The Quintopus comes to mind...it was an SIO expander box that connected to one or two computers and provided an array of SIO ports for connecting multiple devices simultaneously, including "dead-end" devices such as modems and cassette recorders. Switches enabled one or the other computer to access the chain. Both could be enabled, but this only worked as long as the computers didn't try to access the same device at the same time.

 

Antic review here: http://www.atarimagazines.com/v8n1/8bitpowertools.html

 

Quite a rare beast today, I imagine.

[Mathy]

Hello Timbit

 

A2RI is the German abbreviation for A2CI or Automatic Two Computer Interface. Thomas used to sell them, but no longer does. In 20 years Thomas sold five A2RI's, I do not know who bought the other four.

 

I can wholeheartedly recommend the A2RI. Either one of two computers (Atari 8 bit computers with a SIO port) can access the SIO devices connected to the A2RI, but not at the same time. (there is a small delay, before the A2RI will allow the other computer to access the devices)

 

Sincerely

 

Mathy

 

Edit: An expanded version of the A2RI would be great (as long as it's as reliable as the original version). Maybe somebody could add features of the Quintopus to the A2RI.

Edited March 25, 2015 by Mathy

[gilsaluki]

I have used the Quintopus for many years, maybe decades.

[Timbit]

Thank you fellow Atarians. This is good information. Both devices have nice features the other is missing.

 

It appears the Quintopus (made by CSS) may still be available to purchase. I sent an email and will try phoning tomorrow. http://nleaudio.com/css/products.htm

While searching for the Quintopus, I found this page listing some other networking devices such as the Supra MicroNet: http://www.atari8.com/node/207

[MrMartian]

This gives me an idea for an SIO 'switch'... Hmmm...

[Marius]

Mathy: iirc fox-1 has also that zwei rechner interface (2RI)

 

The multiplexer is also a great solution although this one asks for a modification of the Atari.

 

There is one atari with all the hardware connected. That one is called the master. This computer runs the master software and can not be used for user applications.

 

The connected atari's are called slaves. Up to seven slaves can be connected. Those can use all hardware connected to the master.

 

Slaves can also chat with each other.

 

Also pbi hardware can be shared. Which is awesome.

 

Modem can not be shared. But printer can.

[Timbit]

NLE doesn't have any parts left except some XF Single Drive Upgrades. Lance Ringquist said Video61 bought all his Quintopus's. I've sent an email to Video61. Haven't heard back yet.

[Timbit]

I found this comment on the Supra MicroNet from a review in ANALOG magazine:

"Supra provided the Atari community with its first networking product. MicroNet provides nine SIO connectors. Eight are for computers. The ninth goes to whatever peripherals are to be connected. The system served to isolate the computers electrically. It did no software checking, meaning that two or more users attempting to print or save at the same time could trash each others output. (ANL 51, 76)"

Apparently the problem is with writing files. If one computer opens a file for writing, then a second computer accesses a file on that disk, then the first computer writes to their open file, the data can get written into the file the second computer accessed.

Mathy, am I correct in assuming the A2CI/A2RI would have the same problem as the Supra MicroNet?

This problem wouldn't bother me in my use of either device, but I can see it would be a problem in a computer lab or office if people didn't verbally request access before writing files.

I suspect the Handic VIC-Switch would also have similar issues.

[Marius]

Stephen J. Camden still sells the multiplexer.

 

He is providing excellent quality and support.

 

You can find him on this forum.

[ryanmercer]

This gives me an idea for an SIO 'switch'... Hmmm...

 

 

Shouldn't be too complicated, see if there are any opensource KVM's and start there.

[Alfred]

I've got a couple of the Supra Micronet boxes. It worked ok, but there were always problems with too many collisions. It's also a bit slow switching access from one computer to the other. So in the end I went with the multiplexer.

[Marius]

Stephen J. Camden still sells the multiplexer.

 

He is providing excellent quality and support.

 

You can find him on this forum.

 

I meant Stephen J. Carden of course. Probably an autocorrect from my phone, although... what is a Camden?

[flashjazzcat]

Possibly the London borough of Camden.

[IndusGT]

Just spoke with Lance at Video61, he had them but they're all sold out, he can do a production run, but he said that nobody's asked in years, so if you're interested state so in this thread and he'll check it and see if it's feasable

[+orpheuswaking]

I've wanted a Quintopus for awhile now. Preferably the Deluxe Version

[Mathy]

Hello Timbit

 

Apparently the problem is with writing files. If one computer opens a file for writing, then a second computer accesses a file on that disk, then the first computer writes to their open file, the data can get written into the file the second computer accessed.

Mathy, am I correct in assuming the A2CI/A2RI would have the same problem as the Supra MicroNet?

 

I never had any problems with the A2RI. But why/when would two computers access the same file?

 

Sincerely

 

Mathy

[Timbit]

Marty, my understanding is that they were accessing two different files.

[+Larry]

Question -- if you are using some "simple" switching mechanism (rather than a Multiplexer) isn't there an issue connecting two separate +5 VDC SIO power supplies?

 

-Larry

[Mathy]

Hello Timlit

 

Marty, my understanding is that they were accessing two different files.

 

IIRC, the "file accessing" is done purely by the computer. A disk drive only does what the computer tells it to do. How then can data that should be written to one file, be written to another file? That's what I do not understand.

 

Sincerely

 

Mathy

[Kyle22]

Think of it this way: One person is writing about dogs and they call the filename D:MYBOOK.TXT. Person #2 is writing about cats, and they also attempt to write to D:MYBOOK.TXT. The file may end up like this:

 

woof woof woof meow meow woof meow woof woof woof woof meow woof meow meow.

 

This would cause a terrible corruption in a multi-user database application, or multi-line BBS.

 

Most, if not all multi-user systems have a way for the operating system to lock files in use so if another user tries to write, they receive an error message if the file is already open for write by another user.

 

If you're sharing a drive with 2 computers, Atari #1 has no way of knowing if Atari #2 is using a particular sector on the drive.

[Van]

Think of it this way: One person is writing about dogs and they call the filename D:MYBOOK.TXT. Person #2 is writing about cats, and they also attempt to write to D:MYBOOK.TXT. The file may end up like this:

 

woof woof woof meow meow woof meow woof woof woof woof meow woof meow meow.

 

This would cause a terrible corruption in a multi-user database application, or multi-line BBS.

 

Most, if not all multi-user systems have a way for the operating system to lock files in use so if another user tries to write, they receive an error message if the file is already open for write by another user.

 

If you're sharing a drive with 2 computers, Atari #1 has no way of knowing if Atari #2 is using a particular sector on the drive.

After looking over the schema of the A2RI, There is access control on a low level, first come-first serve basis. It's probably true that both Computers shouldn't access the same file at the same time, there isn't any A2RI control for that. How does DOS handles multiple file streams?

With the A2RI given Comp A and Comp B:

If A begins an exchange with a SIO device, it would lock out the access of B from the SIO bus till there is no longer any DI traffic from the SIO device+the RC time const value. I don't know the exact timing here but would guess it's long enough to cover any normal gaps in the serial stream.

Now if B was trying to access the bus during this exchange it would never get a response from the requested SIO device so it would time out and have to retry later.

When A completes this exchange the RC circuit will hold the buss till it times out. At which point both A and B could race to gain control and A may lose, with B starting a new request. How well the drive firmware handles this I don't know.

Yogi

[+David_P]

With both PROCEED and INTERRUPT lines, it should be possible for one Atari to act as an SIO fileserver for at least two others (assuming we don't bit-bang a joystick port for a third one). Ideally, the server computer would have a cartridge or PBI/ECI storage solution.

 

Alternatively, by swapping Rx and Tx lines, plus COMMAND and INTERRUPT lines, two computers could communicate with each other; running a fairly small piece of software, either computer could serve information to its peer.

 

The problem will always boil down to Atari DOSes assuming that they are the sole users of the drives they access. One solution might be a new driver for SpartaDOS for shared drives, where the driver assumes a more intelligent peripheral that will resolve filenames and storage on its own. That way, SpartaDOS would not have to worry about the table of contents or sector map, and the intelligent peripheral can provide errors if multiple users try to write to the same file.

[jblenkle]

I don't know if this would allow you to do it and I have to admit I've never even used the thing, but this has four SIO ports (two of them powered).

 

 

Is it possible to have two or more Atari 8-bit computers share a chain of SIO devices, either simultaneously or in turn? I'm thinking of something like the Handic VIC-Switch for the Commodore 8-bit computers.

I also asked this elsewhere and was pointed to the "Automatic Two Computer Interface" mentioned in the page linked below. It only supports two computers but, based on how it works, I think it could be expanded to support more. You will have to page down a few times to find it: http://www.mathyvannisselroy.nl/special%20stuff.htm

 

At the end of that section, he mentions the A2RI. I'm not sure whether this is the same device but it seems to have the same functionality. This is in German: http://www.mathyvannisselroy.nl/2r_bauan.PDF

 

Does anyone know whether anyone is selling such a device, or know of other options?

 

Thanks.

 

[Mathy]

Hello David

 

Alternatively, by swapping Rx and Tx lines, plus COMMAND and INTERRUPT lines, two computers could communicate with each other; running a fairly small piece of software, either computer could serve information to its peer.

 

Sounds like the GameLinkII/MultiLink interface. (skip the MultiJoy part)

 

Sincerely

 

Mathy