NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/dcom/modems/13364

Internet Message Format | 1992-09-13 | 5.4 KB

Path: sparky!uunet!think.com!camb.com!bruce
From: [e-mail] (Barton F. Bruce)
Newsgroups: comp.dcom.modems
Subject: Re: ZyXEL U1496E+
Message-ID: <[e-mail]>
Date: 14 Sep 92 03:00:28 EDT
References: <[e-mail]> <[e-mail]>
Organization: Cambridge Computer Associates, Inc.
Lines: 98

In article <[e-mail]>, [e-mail] (Paul Bash) writes:
> In article <[e-mail]> [e-mail] (Lam Lai Yin Savio) writes:
>>I am planning to buy a ZyXEL U1496E+ . Although there is no guarantee
>>from ZyXEL that the E+ will be upgradable to v.fast, I would like to
>>hear your opinions on whether it has enough hardware to be upgraded to
>>v.fast? (I know this is a tough question, since v.fast is still at it's
>>early stage of development)
>>
>
> This is kind of a silly question. It really won't matter whether or not it
> _could_ be upgraded unless Zero One decides to do so. Since they haven't

It is hardly silly. Modem history is littered with examples of models
that were supposed to be able to be upgraded but when push came to shove
could not cut it. It was NOT for a lack of trying, just not enough umph
in the box.

I have reasonable faith that ZyXEL would do the upgrade if at all
possible. It may well not be DSP power that kills them, but 68xxx power.
A long time ago I posted some relative performance figures here that
were the results of some informal bench tests Gordon ran one evening. I
probably have them buried some where or I am sure someone else does.

What I remembered they showed, but the difference was so tiny that it
could wrong, was that performance was just starting to roll off when
compressing a very compressible test file using v.42bis on a v.32bis
link. Presumable the limit was the 68xxx processon, not the DSP chips.

The numbers as I recall them were that throughput was 1.45 times the
v.32 (9600) performancxe for that same file rather than the theoretical
1.5. Too small to be valid? Maybe. It was a large file. The 68xxx may
have been running near flat out.

Gordon's comments about not using a faster chip from that family were
simply that the one used was adequate for the current job, and the next
faster one made a BIG difference in price. All this was ages ago in
modem history time. Then almost noone else had v.42bis with v.32bis and
few went to 57.6 let alone faster, and many others lacked power to do
much compression beyond what they did on v.32 modems.

Now many others have caught up power wise, we are in the typical modem
price war again, where companies trot out features to differentiate
themselves from the others while trying to keep their prices up without losing
market share. ZyXEL, too, of course.

So I do worry and care about what they dare or dare not promise. A good
question might be "could the existing modem be easily fitted with a
faster processor in the same family, and would that solve the v.fast
power problem?"

Additionally, of course, 115.2kb would be needed. Even though they might
apparently have no reason today to offer it, I can easily think of a
VALID one. In private line mode, with v.42bis, it is easy to need more
than 57.6, hence that bastard faster one. BUT, how do I get a PC to run
that speed? Or at a hub site for SLIP lines, something like a DEC DS700
terminal server does 16 ports of 115.2kb slip, or 57.6, BUT NOT in
between. If ZyXEL had 115.2, it could be useful today.

If ZyXEL can run at 115.2, they should have, and skipped that weird
speed. But they didn't - what does that tell you?

> A better choice would be to purchase one of the USR or Codex modems that
> _have_ been designated as upgradeable to V.FAST (whatever that means :-).

Well, maybe, or maybe NOT. Can you get the Codex unit for less than
twice the price (street, not list) of the ZyXEL? Probably not. Will the
v.fast ZyXEL that we assume WILL replace the current one IF necessary
probably come to market at about the current ZyXEL price? Given the
modem bang / buck history of recent years that is reasonable to assume.

So I would rather deal with the nice folks at ZyXEL, save 1/2 my money
for now, and maybe spend the same again later for v.fast. In the
meantime, I have money in the bank, neat features ZyXEL has that Codex
does not, and later I would have a second modem - maybe.

If I go with Codex now, I do get a faster modem now, but it costs a LOT
more, I miss some nice ZyXEL features, and I get to deal with a company
that had average POs of over $60K when last I heard. They sell to large
multinational companies that want conservative rock solid stuff globally
identical - price be dammed.

Try to find a Codex sales droid that will talk to you for a pair of
modems and you learn what "small customers be dammed" means! In all
fairness to Codex, they DO intend to support small customers via their
dealers, and via their 1.800 direct sales channel.

If I am way off base here, and if the recent press B/S about Codex's
chief promises new vitality to company is real, and if someone knows
where to get such QUALITY products (yes they are!) at affordable prices,
PLEASE POST or MAIL.

In the meantime let us all continue to enjoy ZyXEL style products and
service and pricing that many find to be a refreshing change in the modem
industry.

What are Codex and USR doing about FAX, Caller-ID, and Voice?

If only ZyXEL would get their own FTP site...