[Simh] Multiple telnet ports in SimH to RSTS/E 9.6

[dundas at caltech.edu]

Happy New Year to all, from Pasadena!  I'm the original author of the VH
simulator and mostly lurk on this list.  I wanted to add my thoughts to a
couple of points that have been brought up.

I too am of the opinion that for Unibus machines, the DH/DM combination
was outstanding.  Great throughput, low system (software) load, extensive
functionality, though high space and power consumption.  For PDP-11s, this
was a great mux.  I used /70s and /45s with these and they could handle
MANY users easily.  Conversely, a DZ could bring any of these to its
knees.  Even worse, when the VAX-11/780s came out, DZ was IIRC the only
supported serial mux and could easily crash the system (VMS V1, V2,
certainly; can't remember if V3 crashed) under load.  Responsiveness was
horrible with DZs under load, even when they didn't cause a crash. 
Ultimately the DHU came out, but it was quite late, as I recall.

Able eventually made a good Unibus device that worked well under VMS. 
DMZ/32?  Can't quite remember.  I do remember the DHDM was good.

For Qbus, in my opinion the DHQ was the best choice, if your software
supported it.  The DHV was a good second choice.  I put together a brief
comparison of the Qbus serial muxes:
<http://dundasclan.org/retro/simh/dzdh-compar.html>

Regarding the VH simulator specifically, I'll note that the simulator
attempts to mimic the DMA (output) capability of the simulated device(s)
by filling a packet with the DMA buffer, rather than one packet per
character (as a DZ/DZV or similar would).

The VH simulator does not attempt to mimic the DH/DM, though it does a
reasonable job at the DHU, getting one pretty close.  Without referring to
documentation, I no longer remember the specific differences, but it
shouldn't be that hard to add whatever logic is necessary for the
remainder of the DH/DM combination.

[Clem, did we meet recently in another context (CENIC/AT&T)?]

John