It seems I have just made a final edition of my blog entry about the VAX-11. It would be great to get some comments. Thank you.
- VCF South West - June 14 - 16, Davidson-Gundy Alumni Center at University of Texas at Dallas
- VCF West - Aug 2 - 3, Computer History Museum, Mountain View, CA
- VCF Midwest - Sept 7 - 8 2024, Schaumburg, IL
- VCF SoCal - Mid February 2025, Location TBD, Southern CA
- VCF East - April 2025, Infoage Museum, Wall NJ
-
Please review our updated Terms and Rules here
A text about the VAX-11
- Thread starter vol.litwr
- Start date
g4ugm
Veteran Member
Don't compare with IBM/370. Totally different market. VAX had nothing like a channel. 370 had no character mode terminals..
Thank you. But would you like to help me to understand this? How might we call the 370 terminals? They are text terminals too, and can emulate some other text terminals via software...
g4ugm
Veteran Member
IBM mainframe screens are normally referred to as 3270 screens or "coax screens". They only transmit data in blocks, so can not emulate other terminals via software. On a VAX when using an editor with a VT100 you type a character, it gets sent to vax, which decides if its a control character which needs acting on or just new text which needs storing in the buffer and echoing back to the user. So for many functions VMS does an IO for each character. Some IO devices can do bulk output.
On a 370 with 3270 display terminals, the screens connect to a 3274 or 3174 terminal controller. If the user types a text character its inserted into a buffer local to the controller and echoed to the screen locally by the controller. You can divide the screen into fields and tell the controller if the fields are numeric or alpha. The users can use the TAB or Newline key to move around the screen. Nothing is sent to the host until the user presses a function key or the enter key. Its only at this point does the host see an IO.
The link between the screen and the controller runs at 1.5Mbits. If the controller is local to the host on a channel it usually runs at 2M/bytes so 16Mbits.
I used to respond to tender documents for IBM and you could tell if the customer had a VAX as there would be also sorts of questions about the overhead of character editing and many would not accept the answer to them was "there is no overhead on the host cpu"...
<<< caveat. the last model controllers could allow 3270 screens to emulate VT102 over TELNET but the emulation is done in the controller. and the controller needs to have a network adaptor.>>
Last edited:
krebizfan
Veteran Member
IBM did put the 9370 in competition with the VAX as low-end mainframes matched up with high-end minis.
g4ugm
Veteran Member
I don't believe IBM put the 9370 up against the VAX. The press did, but I don't think IBM did.
krebizfan
Veteran Member
IBM almost never mentioned a competitor by name but the 9370 announcement letter referred to it as a "super-mini computer." What other successful system qualifies as a "super-mini?"
g4ugm
Veteran Member
I think fans of Pr1me and Data General might well claim they were successful super-mini computers. I feel HP also had boxes that might be super-minis.
As for the 9370 on re-reading the blurb perhaps it was sideways aimed at DEC, Pr1me and DG.
IBM 9370 INFORMATION SYSTEM OVERVIEW
However whilst it has "370" in the name, the 9370 is not a 370 and almost no mainframe person would acknowledge it as such.
Its more of a micro-coded emulation of the 370. - It does not have channels, except as an option, all the peripherals provided as standard were "integrated"
As marketed it does not run MVS IBMs flagship OS, only VM -> you can add extra hardware but then you end up with mainframe prices.
It can have ASCII terminals, but only via an option board that contained memory buffers and made them behave as 3270's to the OS so no EMACS, VI or TECO.
So really a kind of Frankenstein box designed to compete with VAX but without the things that made VAX great...
I think it was one of IBM's few flops and an abject failure to move S/370 down to the mid-range & mini market place.
Its also a bit oddly placed on the timeline of computing. it announced 16 years after the first 370 so not in the same timeframe as the classical 370 derivatives of which in the form of the 4381 and 3090 both of which have real channels and run MVS out of the box and which continued to be offered by IBM. Also six years after the first VAX so IBM was slow to see it as worth competing about. So in summary, its fine to say IBM produced the 9370 an emasculated 370 as a VAX killer and that it failed, but its also important to see it wasn't a part of the mainstream S/370 family.
Its also sad to think that whilst OpenVMS still exists on X64, IBM still make "Z" machines which run modern descendants of the S360 and S/370 operating systems. You can still run problem state object code from a S/360 on the latest Z box., and many customers do this. You can't do that with VAX code without an emulator.
p.s. I understand 9370s were often sold into totally inappropriate situations. Almost every salesman had a 9370 in their sales quota and so would offer one at attractive terms to any customer the felt would buy one.
Last edited:
g4ugm
Veteran Member
Intersting comparison here:-
Back in "za ole days" when there was something called "The Bell System" we did a lot of our daily work on systems called COSMOS, TIRKS, BCAS, and so on, that we accessed on Teletype serial terminals that must have been connected to IBM mainframes because they ran TN3270 "block mode." At some point in the 1980's we were even pulling pending work, closing it out when complete, and accounting for our time on a system called WFA for Work Force Administration. Still on those block mode terminals. So I guess there was some bridge hardware in between that converted character mode to block mode. When I retired in 2014 we were still using those support systems, but the TN3270 emulation was running on Dell Optiplex desktops running Windows.
g4ugm
Veteran Member
There are lots of ways of getting TN3270 into an IBM Mainframe, but the crux of the matter is that if you want a screen mode terminal, it pretty much has to look like a a 3270 to host application programs, and TN3270 achieves that. It packs a 3270 datastream into a TCIP stream. There were other earlier solutions. I think one of the first came out of Yale university. They produced something called "Yale ASCII" that ran on box called a Series/1. This had lots of RS232 ports and an IBM channel adaptor. You could configure terminal code mappings and it looked like a 3270 controller to the host. IBM then produced its own box called the 7171 which was cheaper. The same idea as the Yale ASCII box but pre-packed. Later they produced a a card call an "ASCII Emulation Adaptor" that fitted into a 3174 controller. and enabled ASCII screens and printers to be connected to a channel.
On the 9370 there were cards for 3270 screen mode terminals, or ASCII screens, but just as with the solutions above the ASCII screens must be mapped to 3270 data streams, so no character mode, just block mode.
Many thanks for this interesting document. However it claims that the VAX-II/780 is the 4361 Digital equivalent. So DEC compared the VAX with the 370.
I also remember when some people wrote about the VAX and the 370 that they used for the same job tasks...BTW I thought that the fullscreen mode of CMS looks like a terminal emulator. It seems it was a result of my ignorance.
g4ugm
Veteran Member
Well of course marketing had to compare, but they do so with an agenda, to convince the the world their machine is better. For me the VAX offered a totally different experience. It was all about end user computing, you had a wide choice of small machines, and DEC didn't throw a wobbler when you plugged something foreign in. Designed to be operated and managed by staff with limited technical knowledge.Many thanks for this interesting document. However it claims that the VAX-II/780 is the 4361 Digital equivalent. So DEC compared the VAX with the 370.
BTW I thought that the fullscreen mode of CMS looks like a terminal emulator. It seems it was a result of my ignorance.
On the other hand the IBM machines were more corporate beasts, run and managed by central IT, bought by an IT specialist who "never got the sack for buying IBM". They were marketed as "departmental machines" but I am not sure they were deployed that way.
I suspect I am in a bit of an odd position, as I worked for the Natural Environment Research Council as a systems programmer. We had both IBM 4381 and 9370 running VM and VAX running VMS. The logic behind this was that we were moving to a world where we would write less software and buy more in, and having VAX and 370 gave us the biggest chance of finding something that worked on both. It was a bit like Linux/Windows. Generally we had VAX and IBM fan boys, but everybody got their work done.....
Last edited:
g4ugm
Veteran Member
Actually I still have both, an IBM P390 which is a Microchannel server, and runs MVS and VM. I also have several MicroVax, mostly tabletop models....
You helped me a lot with your Vaxen (MicroVAX 3100 model 85, VAX 4000 VLC) some time ago. Thank you very much again for this. Would you like to run my π-spigot calculator on your P390 under VM/SP? You know I gather results for 100, 1000 and 3000 digits. It would also be great to get results from your P390 under MS-DOS.
Last edited:
g4ugm
Veteran Member
I will when I switch it on next, but this may be some time...