🚀 go-pugleaf

>I was under the impression that Digiex was backhauling traffic between
>wire centers on voice grade FX circuits. I was wrong and shouldn't
>have pointed the finger at Digex, but I would still dispute any
>argument against my analogy.  The "bigger roads" are available to the
>residential (dial-up?) user (switched 56, ISDN, etc), they're just
>more expensive.

We buy "phone lines" from the phone company.  They only come in one 
flavor; you can't buy "good" or "medium" grade lines, they are just
voice grade lines.  If we use FX (foreign Exchange) circuits or 
RCF (Remote Call Forwarding) or SR (Switched Re-direct) lines, they 
are all the same from the stand point of the bandwidth that they are
supposed to carry - they are voice grade lines. 

There is no guarantee as to the capability of voice grade lines - they
are not designed to carry data encoded as tones at a certain rate.  
They are designed to carry the human voice, and in fact the signals 
are shaped to do that as well as they can.  The only assurances that 
the telephone company offers on data for these lines is 1) 2400 baud
modem traffic, and 2) "your fax machine will work over them".  In 
PRACTICE they will try to help if your data lines can't do 9600 or
14.4k or thereabouts, because there has been an effort at promoting 
the awareness of the phone company people that customers really do 
use the lines for modems (Digex's own Jim McDonald, who came to us
from Bell Atlantic, was one of the main drivers of that initiative 
within BA).  Also, if your fax machine can not receive faxes, then
they get onto that because it is a clear case of the machine "not
working".

The telephone company will not, however, give a damn about whether 
you can connect at 24k vs. 28.8k   There are no assurances given 
at all about high speed data (and that is considered to be high speed
data to them) over voice lines.  Period.  The law backs them up on 
this absolutely.  Their own self interest may back them up on it 
too, if they are looking to have people who want data services 
be forced to BUY data services (ISDN, etc.) which they are...

>Theoretically, the real problem here is the development and
>advancement of analog modem technology. It seems as if we're up
>against the bandwidth limits of what's capable via analog POTS
>circuits, and to be fair, I don't think that's Digex's or Bell
>Atlantic's fault.

The modem manufacturers are under a lot of presure from the markets
to build faster modems.  The problem is with the public perception
(which is driven by advertising) that these modems which "are able to
operate at 28.8k" should actually be expected to achieve that level of
performance most or all of the time.  Modems back down in speed to 
compensate for different line charactoristics.  Consumers mistake 
this as the modems backing down because something is WRONG and that 
someone should FIX IT.  That is not the intent of the manufacturers.
The guys who build the modems don't expect them to run at 100% all 
of the time or they wouldn't have such complex and involved speed
back-down facilities.  They would just runl at 28.8k or nothing.

>Clearly, BA (or any new competitor to the market) is
>not going to install new miles of copper (probably obsolete beyond
>analog 28.8 anyway) to serve ever increasing bandwith requirements --
>that would be rediculous. So I'd have to argue that the previous
>analogy holds, but actually applies to us as end users rather than
>Digex, as the ones who are trying to force BIG trucks down little
>roads.

The problem, most likely, is this.  The new equipment, which is 
being installed more and more, is focusing on carrying voice.  All 
the old stuff did that too, but it's margins for error were much 
larger, and it could carry a bit MORE than just voice.  A real
analog phone line carries frequencies up to 4.5 khz, although 
the attenuation (decrese in loudness for the non propeller heads
in the audience) at 4khz and above is pretty bad.  Still, there is 
a signal there, and while people might not be able to make use of
it, modems can.  Strictly speaking though (great pun) voice circuits
only NEED to carry up to about 3.5 khz in order to be perfectly 
fine voice quality; this is tuned for the human ear remember.  What
this means is that when they are building equipment NOW, they have 
the technology to be more careful, to "waste" less bandwidth and to 
deliver those voice calls at less cost.

This is great, unless you were planning on using that "waste" 
bandwidth with a modem.  For example, the 8th band in the v.34
definition, which operates at 3750 hz; that one is hosed if the 
equipment cuts things off at 3500 hz.  That is what the SLC96's 
do, and that is why it is not likely (read that as impossible) 
that you are going to get 28.8k through them.

Now, the rub is, nothing is "broken" from the technical, letter 
of the law, kind of point of view.  There might be a case of 
"unequal access" being provided, and that one is being looked 
at, but that is a different kind of question really.

>So the underlying gripe is really one of price and who should pay. As
>technology has evolved, we as residential users have become accustomed
>to using low cost voice grade service as a means for low cost, low
>speed data transmission (which it was apparently never designed for).

That is the phone companies position, absolutely.

>From now on, if *we* users want increased bandwith (access to the
>bigger roads), I guess we'll have to start paying our fair share for
>it in the form of data circuits and data transmission equipment. The
>alternative of course is to live with the existing speed limitations
>and complain, complain, complain.
>
>I guess the only  ones left to point the finger at are the modem
>manufacturers, they new the limitations, didn't they?

Well, as I said above, they are under a lot of presure...

Peace, 

Doug