[Simh] Additional Interfaces for the VAX780 simulator

[Ethan Dicks]

On 4/14/09, Hölscher <hoelscher-kirchbrak at freenet.de> wrote:
> > On 4/14/09, Hölscher <hoelscher-kirchbrak at freenet.de> wrote:
>  >> I would like to have two (or more) additional interfaces for the VAX780
>  >> simulator:
>  >>
>  >>  DMC-11 (for DDCMP - DECnet, would be useful for the PDP-11 as well)
>
>  > How would you handle the sync-serial nature of the external interface?

> For a start I would be satisfied if I could connect two or more simh simulators
> (all on the same host), so there would be no hardware issues.

Sure... you could use sockets or such for that.

>  Later of course I'd like to connect to the outside world as well.
>  I've got absolutely no clue how to do that yet.
>  I hope for intelligent ideas / suggestions from other people :-))

Well... as I mentioned, the serial hardware on a real DMC-11 (or any
DDCMP-capable interface) is entirely unlike serial ports on modern
machines (or most ancient machines for that matter).  WAN technologies
(sync serial stuff) have always been substantially less common than
local stuff; these days, it's mostly been relegated to appliances like
Cisco routers and such rather than directly, as it was done 20 years
ago.

>  Have you got any idea how difficult the simulation of a CI780 would be?

None.  I never used any CI gear back in the day.

>  no, that's most probably out of reach - but wait, wasn't there a PCI-CI card for Alphas?

Perhaps.  That stuff is all after my time.  If so, then perhaps one
could support that through simh.  Without driver support, perhaps the
easiest place to start would be support simh on an Alpha, but plumb
the PCI-CI card to simh, not to the hosting CPU.

>  Clusters have always been a specialty of VMS and I'd really like to have a simulated CI one (VMS V3 / V4) at least.
>  I already have a real VMS V4.6 LAVC running with a MicroVAX II as boot and disk server and two VAXstation 2000s and another MicroVAX II as satellites,
>  really nice, but I want it older & bigger.

That's a part of VMS history that I never had experience with in the
1980s and early 1990s.  We had several standalone boxes with multiple
DDCMP and HASP links for a sync serial point-to-point network (among
other gear, we had a rack shelf full of sync modem eliminators just to
test our product, so it was always easy to find a spare to link two
machines via DDCMP).  For our MicroVAXen, we used sneakernet and
Kermit over the async ports.

>  > ...what about async DDCMP?  Are you
>  > working in an environment where that's not possible?
>
>
> I'm a little crazy about OLD VMS versions - and there was no async DDCMP before V4.

True.  My experience is mostly with V4 and V5.  I started in the late
V3 days, but I have mostly user-level experience from that time, not
developer or sysadmin experience.

>  Furthermore most PDP-11 OSes do not support async DDCMP either, and I'd like
> to connect into that direction, too.

Gotcha.  That does speak to the need for sync DDCMP, which is why I asked.

>  I must admit it is all for fun, so calling that "working" does not quite hit the point ;-)

Fair enough.  In my case, all of this is for fun, too.  I haven't made
a dollar from VMS since 2003, and that was all modernish Alpha stuff
w/SAN, tape robot, etc.  Prior to that, it was 10 years earlier when I
used all of the old 11/7xx stuff and standalone MicroVAXen.

Right now, I'd just be happy to replicate an approximation of that
1980s-1990s environment - an 11/750 w/8MB, many serial ports (56?
Enough for plenty of interactive users), 500-800MB of disk (to hold
the savesets I still have), tape, floppy (yes... we had a real RUX50
so we could cut diskettes for customers), line printer.  The 11/780
implementation in simh is well along the way to what I am after.  I
just need to sit down and get VMS 4.5 or VMS 4.7 up in that
environment.  Later, I'd like to play with VMS 3.4 or VMS 3.6 for
nostalgia's sake (and to see how much "better" 4.0 was all over
again).

-ethan