*BSD News Article 43253


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!agate!news.mindlink.net!giant!a09878
From: a09878@giant.rsoft.bc.ca (Curt Sampson)
Newsgroups: comp.unix.bsd
Subject: Re: flat rates for Internet/phone
Date: 8 Mar 1995 03:56:49 GMT
Organization: MIND LINK! Communications Corp., Langley, BC, Canada
Lines: 83
Message-ID: <3jj9u1$b0g@deep.rsoft.bc.ca>
References: <D3s19v.4M7@pe1chl.ampr.org> <3jg3g0$n0p@park.uvsc.edu> <3jiddj$5e7@deep.rsoft.bc.ca> <3jio8a$79b@GRAPEVINE.LCS.MIT.EDU>
NNTP-Posting-Host: giant.mindlink.net

In article <3jio8a$79b@GRAPEVINE.LCS.MIT.EDU>,
Garrett A. Wollman <wollman@ginger.lcs.mit.edu> wrote:

>These devices are called ``line cards'' and you actually do pay for
>one on each line.  Both of these things are now done in the same DSP
>chips that are responsible for the A/D and D/A conversion.

I suppose this could be the case. Have you actually checked the
manuals for proper BX switches and verified this? I'm going on the
design of the AT&T Definity G3 switches which are a) designed for
small systems (with less than about 8000 telephones) and b) tend to
use a fair number of digital handsets. None of the analog cards
available have any tone decoding or dialtone ability.

Even if this is the case on newer switches, though, we have to keep in
mind that it could easily be twenty years or more before telephone
companies have ammortised their older switches and get rid of them.
(Our local telco has to ammortise its switches over about thirty or
thirty-five years, from what I've heard.)

>>Given that telephone calls are a circuit-based application, why on
>>earth would the telco want to move to packet switching? Circuit
>>switching seems to me ideally suited to dial-up telephone networks.
>>What advantages would packet switching offer?
>
>The ability to carry something other than not-very-profitable phone
>calls.  Most existing models of CBR voice/VBR data integration either
>over- or under-utilize the available bandwidth; replacing the
>guaranteed virtual circuit service used by most phone calls with a
>predictive packet-based one would provide better integration and line
>utilization in such an environment.

Let's make an assumption here: for the next fifty years or so, most people
are going to want to use traditional analog handsets for their residential
telephone service becuase they are much, much cheaper than digital
handsets. (They're going to remain that way becase we still haven't got an
extremely common digital standard yet. Perhaps one day we will be able to
buy ISDN telephones for $10 or so, but I suspect that day is many years
off.)

In this case it makes sense to continue to use the current circut-based
system for quite a while. All the equipment is available, it's
cheaper, and it provides better behaviour in the event of high loads.
(People just can't get start a connection, rather than having current
connections deteriorate and fall apart.)

So then we could continue to use the current system for regular
telephone connectivity, and if someone wants packet connectivty the
telco could run a second line, and instead of plugging that into the
switch in the CO the could plug it into their packet-switch instead,
and avoid any use at all of their switching apparatus.

As a matter of fact, I've heard rumours that our local TelCo is
working on just such a thing. They would provide a 128K copper line to
the CO, which would then go into their packet switch/router/whatever
and connect the user directly to the Internet. This would be a 24-hour
permanent connection. The equivalant of the terminal adapter that the
user would need would be quite cheap, since it wouldn't need any of
the ISDN signalling and control circuitry that an ISDN TA needs. The
price would be, apparently, $50 to $75 per month for residential
users.

Most people wouldn't care about this. ("I've already got a telephone
line and a modem; why would I want to pay $50 per month instead of $20
for unlimited internet access?") The more diehard people, like you and
me, would jump at the chance of a 128K permanent connection for the
paltry price of $75 per month, and trade our second telephone lines
for this. This would get rid of the worst abusers of the circuit
switched system, leaving it free to service the rest of the people at
reasonable levels for a circuit switched system.

I can imagine a day when everybody would just be using packet switched
systems. But that's not going to come until everybody can get the
equipment to use such a system as cheaply (or nearly as cheaply) as
our current connection-based telephone/communications system, and a
lot of connection-based equipment has been written off.

cjs
--
Curt Sampson  a09878@giant.rsoft.bc.ca		Opinions are mine,
Fluor Daniel Wright, Ltd. 604 488 2226		not Fluor Daniel's.
1075 W. Georgia Street
Vancouver, B.C., V6E 4M7	 	De gustibus, aut bene aut nihil.