[email protected] (Edward Jaffe) writes:
> You've described the old CUT-mode interface. Somewhere around the
> early 1980s, DFT mode was introduced. It does not encode the data, use
> a screen to send it, or any of that. It simply wraps the binary data
> in a 3270 structured field envelope and sends it via the WSF
> (write-structured-field) command. Any 3270 emulator that still uses
> the CUT-mode interface for IND$FILE should be tossed on the trash heap
> of history...

note that tcp/ip ftp ... typically does full-duplex pacing trying to
maximize packets in flight (aka peer-to-peer networking). 3270 dumb
terminal emulation tends to use half-duplex end-to-end serialization.

tcp/ip "slow start" was introduced in the latter half of 80s as
congestion control ... start with small number of packets in flight and
slowly increase the number of packets in flight ... until get lost
packet indication and then back-off number of packets. 
http://en.wikipedia.org/wiki/Slow-start

in the mid-80s, there was adaptive rate-based pacing as part of my
internal high-speed data transport project ... some past posts
http://www.garlic.com/~lynn/subnetwork.html#hsdt
old paper I did on adaptive rate-based pacing for XTP protocol (i was on
xtp technical advisory board over extreme objections from communication
group)
http://www.garlic.com/~lynn/xtprate.html

more recent there have been efforts with internet2 and next generation
internet with rate-based pacing ... that shows significant increased
throughput ... a subject I will periodically pontificate on. random
rate-based pacing reference here
https://lists.internet2.edu/sympa/arc/transport/2005-02/msg00004.html

change from 3272/3277 to 3274/3278 move a lot of electronics from the
terminal back into the shared control unit (to cut down manufactoring
costs). as a result there was enormous increase in protocol chatter over
the coax for 3274/3278 (as well as lots of increase in latency). later
with 3277 emulation cards, they get three times the upload/download
throughput of 3278 emulation cards.

early 80s there was lots of work on showing human productivity for
.2second response ... but 3274/3278 made that responsible ... past
post with analysis of 3272/3277 and 3274/3278 from early 80s.
http://www.garlic.com/~lynn/2001m.html#19 3270 protocol

as referenced, none of the mvs/tso systems were even in the running with
best possible of one second.

somebody published a report that their internal interactive operation
was best in the company with quarter-second response.  I complained that
I had operations on the west coast with .11 response (i eventually got
back response that they could claim anything they wanted).  3272/3277
had hardware latency of .086 ... so to meet requirement of .2second
required system response no more than .114sec. 3274/3278 best case
hardware latency was .283sec ... but more typical was .530sec (making
.2sec response impossible).

we complained to the 3274/3278 product people ... and eventually they
came back that 3274/3278 was for data entry and not intended for
interactive computing,

from recent discussion
http://www.garlic.com/~lynn/2012m.html#15 cp76, vm370, etc

from ibm jargon:

bad response - n. A delay in the response time to a trivial request of
a computer that is longer than two tenths of one second. In the 1970s,
IBM 3277 display terminals attached to quite small System/360 machines
could service up to 19 interruptions every second from a user I
measured it myself. Today, this kind of response time is considered
impossible or unachievable, even though work by Doherty, Thadhani, and
others has shown that human productivity and satisfaction are almost
linearly inversely proportional to computer response time. It is hoped
(but not expected) that the definition of Bad Response will drop below
one tenth of a second by 1990.

... snip ...

-- 
virtualization experience starting Jan1968, online at home since Mar1970

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to