Subject: | |
From: | |
Reply To: | |
Date: | Wed, 10 Feb 1999 13:25:55 EST |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Gavin writes:
> No ENQ/ACK handshaking on any version of MPE/XL or MPE/iX that I'm aware
> of. DTCs have always used pure XON/XOFF handshaking I believe, and TCP
> based connections (VT, telnet) don't use in-band flow control at all (apart
> from the read trigger stuff).
>
> Someone correct me if I'm wrong. Ross?
Gavin is absolutely correct about ENQ/ACK not being supported on MPE/XL or
MPE/iX. That was a pacing structure that was used only on the Classic
machines, designed primarily to see if the terminal was still connected to the
HP3000. It disappeared with the advent of the RISC-based devices (although
there was no correlation between the two; they just occurred together).
And Gavin is currently correct about Telnet not using the XON/OFF flow control
protocol, but that's only currently. XON/XOFF (Telnet option 33) has just
recently been added by Jeff Bandle of CSY's Networking Lab to MPE/iX and will
soon be appearing on an HP3000 near you.
What is debatable in Gavin's comment (and this is not meant to be in any way
derogatory to Gavin) is to call XON/XOFF "in-band flow control." XON/XOFF
signals are always transmitted in the reverse channel, not the primary, thus
there is never any confusion as to what is data and what is control. Even on a
full-duplex line, communication with the HP3000 is always just one-way (and
then the other). It is never "full-duplex" on both halves of the line,
simultaneously.
Wirt "who's name sounds a lot like Ross, when heard at sufficient distance"
Atmar
|
|
|