W8SDZ@WSMR-SIMTEL20.ARMY.MIL (Keith Petersen) (01/07/89)
More quotes from Toby Nixon of Hayes... presented "as-is" for
informational value. I am not the author.
--Keith Petersen <Info-Modems-Request@WSMR-Simtel20.Army.Mil>
-----
Date: 01-05-89 (17:26)
From: TOBY NIXON
Subj: CCITT
The CCITT has a couple of active projects related to modulation.
First, they are planning on extending V.32 to higher speeds such as
12000 and 14400 bit/s. With V.42 error control and data compression,
we're talking about transfers at 43000bps and above, full duplex.
Second, they are continuing work on the asymmetrical modem standard.
This is a synchronous, non-error-control modem that transmits in both
directions simultaneously but at different speeds. Currently the
project is somewhat confused, since the proposal contains multiple
different and incompatible schemes (one similar to the _modulation_ in
the USR Courier HST, another similar to the _modulation_ in the
current Telebit Trailblazer, but in both cases differening in
IMPORTANT areas).
It also includes provision for a fast-turnaround half-duplex ping-pong
operating mode similar to the V-series Smartmodem 9600, but, again,
different in some respects. NO WORK has been done on extending this
to include async-to-sync conversion, error control, and data
compression, to get it even up to the level of the present V-series
9600.
Personally, I think the real "battleground" is in the area of V.32
_implementation_: who can do it best and cheapest. V.32, even at
9600, with V.42 and data compression will give you very impressive
performance. The arguments over the tradeoffs and compromises
represented by each of the various techniques for "pseudo" full-
duplex high-speed communications seem to be intractable, and we'll
probably end up with multiple standards (ick) or no standard at all
(except, of course, V.32 and the new extended V.32, which ought to be
good enough!)
Hayes evaluated all of the different techniques (asymmetrical,
multicarrier, dynamic duplexing, etc) before choosing the one we did
-- ping-pong -- because it gave the best all-around performance in a
variety of applications, very quick echoplex responsiveness, good
noise resistance, etc. We give up a very small bit in peak one-way
transmission capability for great advantages in bidirectional data
transfer. Other manufacturers look at the same techniques and the
expected mix of applications among their users, and some chose similar
solutions and others chose different solutions.
Rambling as well, my point is that the BEST all-around solution is a
full-duplex modem, but they've been too expensive. Our quickest and
easiest solution was to provide different capabilities at a lower
price; now we're working hard on providing those full-duplex
capabilities at a similar low price.
The solution to the high-speed compatibiility problem is for everyone
to implement the _standard_ (i.e., V.32 with V.42), but we don't want
to lose a great portion of our market because of the current high
price. Those manufacturers who get high-quality, reasonably-priced,
full-featured products on the market quickly will reap the rewards;
winning this "battle" in the standards committee is not, in reality,
going to benefit anybody.
-- Toby