Eero Laurila <[log in to unmask]> wrote:
>With VT, the same logon string is going to be 2 TCP packets over the net.
 
More like four, or maybe three if you can type that logon string
really fast.  Those TCP acknowledgements need to be sent somehow, and
you probably won't be able to piggyback more than one of them.  But
what's a packet or two between friendly hosts?
 
>If line mode can be used -or- if both VT and telnet are used to interface
>with an application doing 1 byte reads -- I would expect their performances
>to be close to each other and that's the best case for telnet.  I don't
>expect block-mode to work through telnet and VT is going to continue to
>have a big performance advantage there.
 
With the WIN/TCP for MPE/V Telnet server, block mode works just fine,
thank you.  (The client is another matter, but its hangup is mostly in
classic MPE terminal I/O.)  And guess what -- block mode is the other
place Telnet performance can approach VT.
 
Even when you're in Telnet's "character" (as opposed to Telnet "line"
mode), Telnet doesn't require that you send each character in its own
individual packet; if you can get a bunch of characters ready to go in
a reasonably short time you can do one big network write and get them
all out there in one (or more) large packets -- not too unlike a VT
read completion in terms of number of packets.  As it happens, HP
terminals (and emulators thereof) are pretty good at getting a bunch
of characters ready to go in a reasonably short time when they're in
block mode, so the real requirement is that you not have a boneheaded
Telnet client that insists on sending each character wrapped in its
own packet.
 
VT has two advantages over Telnet.  First, it forces the client to do
presentation details like echoing and line editing for edited-mode
reads, so that you don't have to waste your 3000 and/or network
dealing with it.  As a side effect it cuts down on network traffic.
Line-mode Telnet can do this too, but since line-mode is an option
(and not a terribly well-supported one as far as I know) any Telnet
server is likely to require local support for line editing facilities.
 
(Hmmm.  I seem to have forgotten enough about line-mode Telnet that
I'm really not sure whether all the little twisty passages of MPE line
editing could actually fit into it.)
 
The other is that VT (so far as I know) keeps its control information
somewhat more out-of-band than Telnet, and so it doesn't have to make
a pass through the incoming and outgoing data looking for special
bytes (as does Telnet).
 
-Frank McConnell, not speaking for The Wollongong Group
 <[log in to unmask]>