3 Jan
2022
3 Jan
'22
4:13 p.m.
On Jan 2, 2022, at 9:14 PM, Grant Taylor via cctalk
<cctalk at classiccmp.org> wrote:
On 1/2/22 6:59 PM, ben via cctalk wrote:
That was the whole point of LAT. At the time almost all serial terminals were connected by
DZ-style interfaces, so 1 interrupt when you typed a character and a second interrupt
after the character was echoed. LAT created a virtual circuit between the terminal server
and a specific host. The slot layer ran over the virtual circuit between a logical
terminal on the host and a physical terminal on the server. Data messages at the virtual
circuit layer were limited to no more frequent than a terminal server parameter ( default
80mS). With appropriate optimizations and data availability it was possible to shrink the
interrupt load on the host to 1-2 interrupts/ethernet packet (say around 1400 characters).
There were a lot of internal religious wars around the use of CTERM/Telnet/LAT.
After finishing DECnet-Ultrix I had a couple of weeks of spare time and implemented a
prototype Ultrix-LAT - one of the good things that came out of having a MicroVAX I in my
office!
John.
But can the Pi handle a gazillion students all
time sharing at once @ 2400?
I think that will depend on how you connect the serial terminals.
I know that it's possible to establish network connectivity to serial terminal
servers. I don't know how many of those terminal servers can be connected to SimH.
It may be better to use LAT between terminal servers and VMS running in the emulated VAX.
How long was
the VAX timesharing era as I suspect networked PC's come out soon after that.
Good question.
--
Grant. . . .
unix || die