Index Home About Blog
From: decvax!utzoo!henry@ucbvax.Berkeley.EDU
Date: Thu, 1 Oct 87 07:15:28 edt
Subject: Re: "separate data network" is silly speculation

> ... [ISDN] ... The resultant network would provide
> 64kbps circuit switching and voice (digitized at 64kbps) for about the
> same price, over the same lines, and throw in access to X.25 as well.

I feel compelled to inject a cynical comment here.  Anyone who thinks that
ISDN will be as cheap as analog voice service any time before the year 2100
is dreaming.  Until such time as ISDN becomes the default for POTS (Plain
Old Telephone Service), the phone companies will have enormous incentive
to charge all the traffic will bear, to help keep the POTS price down.

				Henry Spencer @ U of Toronto Zoology

From: (Fred R. Goldstein)
Subject: Re: DOSBS on PBX?
Date: 2 Jul 1996 04:31:23 GMT

In article <1996Jun30.061834.1@eisner>, says...

>I am specifically interested in using a SX200D that has a T1
>from telco. NO I don't want to pay MITEL to make it a LITE
>system and then pay for their $7500 LITE PRI box only to
>be ripped off BUSINESS MESSAGE UNITS from home for telecommuting,
>I just want to dial in in "VOICE BEARER" mode and come in on a
>normal T1 from the CO and into BRI connected TAs on the back side 
>of the PBX.

Ten years ago somebody at Mitel told me the industry's dirty little
(not so) secret -- they don't build BRI sets because  they want to
lock you in via proprietary desktop sets.

Anyway, all you need is an "R-interface" data system.  Take a router
with 56k sync serial ports, for instance, and  hook them to the 56k
sync serial applique that most PBXs have.  Make a DOSBS call into a
DID T1 trunk and it'll come out of the applique at 56 kbps.

I did this for a while.  I used DOSBS (Gandalf) into an ISDN CO to
call a non-ISDN CO into a non-ISDN SL-100 PBX with a Datapath (56k) 
unit on it; that fed a Gandalf 56k bridge port.  It even worked between
PBXs over T1 trunks, using call forwarding!

Now if that isn't a good use of DOSBS that *isn't* just a tariff hack,
then what is?
Fred R. Goldstein   k1io   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn
Subject: Re: How to be billed as voice, over ISDN????
Date: 8 Jul 1996 02:44:50 GMT

In article <4rkboe$>, says...

>Correct.  The customary term for this sort of "cheating" (:-) is DOSBS:
>Data Over Speech Bearer Service.  The terminal equipment at the other
>end of the link must be both equipped and configured for it,and if you
>get "lucky" and such a call gets routed over an analog trunk in the
>middle somewhere... buh-bye.

The odds of hitting an analog trunk on an intra-LATA call between two
ISDN switches in the United States is approximately zero.  There are 
no analog trunks, except to a few really remote outbacks which don't have

DOSBS is tricky inter-LATA where IXCs customarily use echo cancellers,
etc., which need to be suppressed; that's not so reliable.
Fred R. Goldstein   k1io   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn,dc.general,,md.general,
Subject: Re: ISDN MD--Last Chance
Date: Tue, 16 Jul 1996 10:59:49 EST

In article <4sel8g$> (Doug
Humphrey) writes:

>In a nut shell, ISDN burns time on the ESS switches; ADSL has
>nothing whatsoever to do with the phone switches (the data part
>of it doesn't; the voice phone line part does of course).  This
>means that an ADSL data connection is not seen on the ESS at 
>all.  Cost for an ADSL port - very cheap; cost for a port on
>an ESS - very expensive.  Cost for usage per minute on ADSL - 
>nonexistant (interoffice costs are covered below); cost per
>minute for ESS Switch Module "time" - expensive (an ESS can 
>only switch so many simultanious calls for each Switch Module
>that it has installed - switchmods are very expensive.  

How do you define  "expensive"?  Have you been reading too many BA rate case 
pleadings lately?  Remember, in this business, "under oath" means nothing.

Lucent now offers a 36 CCS line side switch module.  They think the 5ESS is 
just ducky for internet access applications.  Even on the old ones, the 
incremental cost of usage, when you added it all up, was a fraction of a penny 
a minute.  Not zero, but nothing like you make it sound.  Remember that each
hundred grand of switch can carry a *lot* of call minutes in its lifetime.

ADSL fabric isn't even in volume production yet.  The stuff that's best known 
was designed for video broadcast.  See below.

>In ISDN, interoffice traffic moves as 56k/64k DS0 channels - 
>one for 56k or 64k ISDN, two for 128k.  ADSL moves as IP 
>(in the case of an Internet connection) packets - 
>10 to 50 times as efficient usage of the interoffice bandwidth. 

A ha!  Here's your fallacy. ISDN, you see, travels just like a modem does, 
over DS0 channels.  Now in MD, flat rate analog (modem) lines cost something 
like $15 month.  ISDN is just a digital interface to the exact same network.  
The line card is different and the D channel eats packet handler instead of 
tone register time, but otherwise a call's a call.  Shall we shut down the 
POTS network because it's too expensive to provide all these DS0 channels?  
Well, at their current prices, BA is almost grotesquely profitable. (Not that 
they make a bundle on resi lines per se, but they don't lose their shirts 
either while waiting for toll minutes to fill their coffers.)  ISDN is just a 
different line-side connection.

Now what's this ADSL "IP packets" stuff?  The model you imply is that of the 
telephone company, who of course is the only provider of ADSL, also being the  
sole provider of IP service!  You want them to monopolize the ISP business the 
way they monopolize dialtone?  You want BA's law department, the folks who do 
the pricing portion of the marketing function, to start making up prices for 
IP service?  You want BA to use its monopoly cudgel to drive the free market 
ISPs out of business?  Already BA is asking the FCC to allow them to surcharge 
ISPs a couple of cents a minute on INCOMING calls.  Now you want to deliver 
the final blow to the residential Internet access industry?  Cute.  At least 
with modems and ISDN (both the same DS0 resources, remember), the telco 
subscriber has the choice of where to dial in to.  It might be an ISP, it 
might be the office LAN (inside the firewall -- just try that with your
telco-monopoly ADSL-based Internet!), it might be an information service
provider.  Cable modems share this flaw with ADSL, but then both media were
designed for video broadcast entertainment delivery, not two-way 

>The improvements that BA and others have made to switches are
>to allow existing switches to be able to do ISDN.  This is fine
>when one in 1000 houses is going to order ISDN.  If one in 10
>or one in 4 order it, then they basically have to order additional
>ESS's, very very expensive (billions of dollars potentially), 
>because they are not just adding functionality, they have to 
>add capacity.  ESS systems are sized on an average call duration 
>of 3 minutes (average call for voice).  Data calls tend to average
>10 times that (30 minutes) to 100 times that (300 minutes).  That
>means that a modem call averages the same switch load as 10 to 100 
>voice callers.  A 128k call with ISDN takes 2 switchmod slots, so
>that is 20 to 200 voice call equivs.  

You are obviously not a traffic engineer.  Switches don't give much more than 
a goat's bzadeh about average call duration. They care about aggregate 
traffic, as measured in CCS (or Erlangs). If 10% of lines were ISDN and were 
making on average 3 30-minute calls a day, and 90% were analog and making 20 
4-1/2-minute calls per day, then both would be generating the same traffic and 
the switch wouldn't care.    Now a 2B call (128k) does take two slots, but 
then I don't think 2B+D needs to be provided for the same price as analog. 

Also, the *alternative* to ISDN is *not* telco-monopoly IP-over-ADSL.  It's 
modems.  There are *millions* of modems sold annually in the USA, with a large 
and growing percentage of switch utilization being modem traffic.  Modems are 
slower, so the calls take longer and the overall CCS goes up.

>>BA just doesn't seem interested in letting people at ISDN.

>If BA could make money on it, they would be very 
>happy to let people use it.  ISDN, as it happens, is 
>horribly expensive as a service - it was designed to use
>the voice phone switch for the data path - a mistake that 
>only a phone engineer would make ;-)  Seriously, if they 
>are charging per minute, then they can make money with it 
>or at least cover their tails, because their COSTS are per 
>minute costs (ESS Switch Mod time), not just fixed asset
>costs.  Problem is, when they are forced into "flat rate"
>structures, then the price is very high because they can 
>not afford to get burned.

With a flat rate structure, they must cover *average* costs with the flat 
rate.  AVERAGE usage of flat rate resi ISDN lines, where they exist, is around 
40 hours/month.  COST to telco of calls is under a penny a minute peak, near 
zero off-peak.  You figure it out.  Or, for the arithmetically challenged, if 
we assume peak-hour cost at .5cpm and off-peak at .1cpm, and 75% of resi 
ISDN is "peak" (because we'll even be so generous as to assume that resi users 
are on switches that peak in the evening), then we have 75% of 40 hours is 30 
hours at 60*.5=30 cph, or  30*$.30=$9/month, plus 6 cents for each of 10 
off-peak hours, total $9.60/month for the average resi user.  This is a lot 
more than the Delaware PUC computed, but a lot less (by around $240) than BA 

Hey, let's be wicked generous.  Let's assume the average resi ISDN flat rate 
user generates 60 hours a month, and it costs a penny a minute at peak.  That 
is still (45*$.6=$27 + 15*.06=.90) under $28/month in actual telco 
usage-related costs, and that includes the extra line modules as usage related.

Fred R. Goldstein   k1io
BBN Corp., Cambridge MA  USA         +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: fgoldstein@bbn.|nospam.|com (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn
Subject: Re: PRI Pricing in Georgia ???
Date: 23 Jun 1997 16:24:28 GMT

In article <>, says...

>        Does anyone know the approximate pricing for PRI ISDN from Bellsouth
>in Georgia ?

I am not making this up.

Last year I wanted to know that.  I had a fax sheet from BS on their PRI
pricing and couldn't figure out exactly which elements applied when -- there
are many of them, and no single "PRI" price.

So I phoned up the ISDN PRI product manager and asked.

He didn't know either.

And while we have *hundreds* of PRIs from Bell South, we still don't know
what the *correct* price is.  But I'm sure they're working on it.

I suspect it'll be around $1400/month in GA, plus or minus $400 or so, but
don't hold me to it.  Unless they've redone the tariff lately, they were
really clever and tariffed a whole bunch of different price elements which
collectively make up a PRI, but didn't know how to put them together.
Fred R. Goldstein   k1io    fgoldstein"at"
BBN Corp., Cambridge MA  USA         +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: fgoldstein@bbn.|nospam.|com (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn
Subject: Re: Always On/Dynamic ISDN (AO/DI)
Date: 24 Jun 1997 18:44:29 GMT

In article <>, says...

>Another issue that will help AO/DI acceptance is if ISPs are
>considered common carriers and have to pay a settlement charge to the
>local telcos.

As you know, that ain't gonna happen.  And if it did, there'd be such blood
on the streets that ISDN would hardly be noticed if it took collateral
damage.  And even then if it did, the ISPs and CLECs would be there to
complete the "doomsday scenario" in which the ISPs all become, or team up
with, CLECs, so that the Bells get NONE of the settlements and the ISPs end
up paying themselves.  Of course if that happens, then under some
mutual-compensation agreements, the Bells would have to pay MORE to the CLECs
to terminate this TOLL traffic than they do for LOCAL traffic....

>Most ISDN equipment can support AO/DI with a software upgrade (some
>devices may not have enough physical memory.)  Seventeen members of
>the 25 VIA (Vendors' ISDN Alliance) members have committed to support
>AO/DI by the end of the year; one has announced their product last
>week, and at least three others will announce soon.  Telenetworks is
>the leading supplier of protocol source code to OEMs.  They are
>supporting AO/DI so that all of their customers (Router manufacturers,
>TA manufacturers, etc.) will be able to support AO/DI soon.

I'm happy that your company, Telenetworks, is on the case.  Of course it's
not just an ISDN stack upgrade; it's the overall routing code, the logic,
etc., that change, and it will probably take a lot of memory, a lot of code,
and most of all a lot of debugging.

>Given all these facts, I believe that we will see the 6 requirements
>that you mention fulfilled.

I don't see the connection.  You tell me that you're working on checklist
item 4 and that Bell South is working on items 1, 2 and 3.  NYNEX is, as
David said, toast.  PacBell still exists as a separate subsidiary of SBC
(which is not the fate that Bell Titanic plans for NYNEX -- it's gut and run)
so they too might get with the program.  But that does nothing for the other
telcos, or all of the other issues.

As I said, a good idea but facing a highly uncertain future.

>>1)  B channel flat-rate usage must be capped so that there is an incentive to
>>disconnect; this could be a "threshold" tariff (e.g., 200 hours) at the telco
>>OR at the ISP end,
>>2)  B channel usage must be cheap enough that users are willing to let it be
>>heuristically controlled; for instance, all-measured rates (when POTS is flat
>>rate) will discourage widespread ISDN acceptance enough to kill AO/DI
>>(again, a threshold rate works),
>>3)  D channel usage (packet charges, not monthly rates) must be either free
>>or nearly-so; current kilosegment rates will absolutely kill AO/DI,
>>4)  low-end remote router products like the Pipeline 75 and Cisco 766 must
>>support this; most don't have D-channel packet at all now,
>>5)  central-site hubs like the MAX, PortMaster and Cisco 5200 must support
>>this; most don't have D-channel packet at all now,
>>6)  ISPs must adopt it and price it reasonably.
>>It is still within the range of realistic possibilities, but given the
>>6-out-of-6 requirements, it's a long way from a sure shot.

Fred R. Goldstein   k1io    fgoldstein"at"
BBN Corp., Cambridge MA  USA         +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: fgoldstein@bbn.|nospam.|com (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn
Subject: Re: Always On/Dynamic ISDN (AO/DI)
Date: 25 Jun 1997 16:02:57 GMT

In article <>,

>>As you know, that ain't gonna happen.
>You're right.  This won't happen if the people in our government are
>as intelligent and honorable as the rest of us.

Not the issue at all.  Govt. responds to political/market forces.  Even the
RBOCs are divided on this, while virtually everyone else opposes it.  Plus
various other reasons not worth going into here....

>But the threat of a settlement charge may make it easier to eliminate
>flat rate.

Not at all.  Flat rate is politically sacred.  Telcos have been trying to get
rid of it for two decades, even as the marginal cost of usage minutes has
plummeted, making flat rate more economically correct than it was 20 years
ago (when usage-sensitive costs were almost a fifth of the whole local pot,
which is higher than today).  Competitive markets will preserve it, not
eliminate it.

>But what if IP gets a workable QOS method and internet voice becomes
>more popular.  Won't ISPs then perform the same service as common

Irrelevant; this is a red herring argument by the RBOCs and especially the LD
resellers who dominate ACTA.  Again I've worked on this issue and don't feel
like rehashing it all right here.

>...There is a fair amount of software "glue" between the three
>main protocols (ISDN, X.25 and MP.)

That's where you get stuck.  (Sorry, couldn't resist.  But in this case it's
true, as well as an old egg-and-elephant joke.)

>As far as hardware, most TAs have
>everything needed except possibly memory.  AO/DI will take a lot of
>memory, possibly as much as 1/2 Mbyte.  But many TAs have enough

Newer TAs have plenty of RAM, but I'm concerned about the Flash that is used
for program store.  Many routers run out of Flash too early; it's costlier
than RAM.

>Note that central site hubs will probably have a PRI, so the X.25 will
>come in on one or more B channels.  The hubs may need X.25, but this
>is still only a memory issue.  The HDLC controllers are there.  And
>the more-expensive central site hubs probably have more memory

Here's where it gets trickiest.  If this following issue is NOT solved, then
it's Dead Meat.  If it IS solved, then it has a chance:

Calls arrive in a hunt group spread across multiple PRI-server boxes (MAX,
Total Control, 5200, PortMaster, etc.).  As calls drop and come back up
"dynamically", they go to an available channel in the hunt group, which could
be on any one of the servers.  Does the X.25 have to be on the same server?
If not, does the X.25 server require the circuit-mode call to always arrive
on the same circuit-mode server?  Or is there (as needed for it to work)
TOTAL flexibility wrt the circuit calls?  Please don't tell me about how BACP
tells the remote WHICH server to dial into; in the real world, each server
doesn't generally have its own phone number.  We have *dozens* of MAXes
serving individual hunt groups, for instance, and I'm aware of ISPs who put
even more into a hunt group than we do.

I suspect that it'll take a fancy combination of AO/DI and L2TP (or whatever
it is these days) to be more practical than a demo.  Again, more delay.

>NyNex may be "toast" but Bell Atlantic still has the same issues in
>the former NyNex areas.  And the same people, more or less.  A couple
>of months ago, NyNex demo'ed AO/DI and they were pushing the
>technology.  I don't know what Bell Atlantic is going to do.

No, they don't have the same issues or people unless they choose to keep
them.  I am not convinced that they will, and I don't think most of the
people in responsible positions at NYNEX expect to be kept, though this is
obviously a sensitive topic... the good ones will probably be at CLECs by
this time next year.  BA's corporate agenda is entirely different WRT ISDN. I
do give the NYNEX people credit for supporting AO/DI, but I doubt they'll
remain to see it implemented here.  (BA's agenda, btw, is spelled
C-e-n-t-r-e-x, not ISDN, just in case anybody asks.)
Fred R. Goldstein   k1io    fgoldstein"at"
BBN Corp., Cambridge MA  USA         +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: fgoldstein@bbn.|nospam.|com (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn
Subject: Re: Nynex wants $3100-10K to install!
Date: 5 Aug 1997 17:36:48 GMT

In article <5s5dki$i54$>,

>This seems very wrong to me.  Can they really ask me to pay some indefinite
>amount to upgrade their equipment to support the service they advertise?  Of
>course arguing with the orders clerk is no good -- is there anything I can
>do about this?

Most telephone companies assume that some small percentage (5-20%) of
subscribers are beyond the 18 kf limit, and base their prices upon the
average cost of simple and complex (>18kf) installations.  Last year's
California rate case had some interesting discussion, for instance.

NYNEX has a relatively high percentage of subscribers beyond the 18kf limit,
due to their own design stupidity (they never turn down an opportunity to use
a costly, lossy long wire when electronics could do the job cheaper).  They
adopted an ISDN policy of "no midspan repeaters".  You got their quote for
the rare exceptions!  The states have say over the policy but have accepted
it so far.

What they don't tell you:  IF they have a SLC-96 type unit within 18 kf of
you, then they will use that as the basis of the measurement and serve ISDN
off of the SLC with no surcharge.  They also don't tell you if they have one,
unless your existing analog service happens to be off of one.  You could be
next door and they wouldn't tell you.  You have to find it out from their
engineers, or from knowing what one looks like (actually, they're usually
buried in funny air-conditioned manholes with power meters on them and often
trucks parked nearby).  Most recent developments have SLCs.  That's your best
Fred R. Goldstein   k1io   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: fgoldstein@bbn.|nospam.|com (Fred R. Goldstein)
Newsgroups: comp.dcom.isdn
Subject: Re: Too far for ISDN
Date: 21 Sep 1997 01:52:25 GMT

In article <>, says...

>Fred, didn't he say they admitted there were half-taps on the line?
>How can you get them to even try and clean up the pair....?

The real answer is that there are many pair on the street, in a big fat
cable, and the specific bridge taps are probably unique to each.  If they
wanted to find a working pair, at that distance, they could.  Their policy is
to search for a NON-working pair so that they can DISqualify as many as

A trick that has been known to work is to place an order for a *business*
ISDN line, under the costlier tariff.  That will tend to qualify more easily,
though the rules are *supposed to* be the same.  Gee, I wonder why.  Then you
cancel it and turn it into residence service.
Fred R. Goldstein   k1io   fgoldstein"at"   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

From: fgoldstein@bbn.NO$ (Fred R. Goldstein)
Newsgroups: nyc.general,comp.dcom.isdn
Subject: Re: NYNEX ISDN billing
Date: 9 Oct 1997 18:52:14 GMT

In article <>,

>Be careful. According to the NYNEX reps that I have spoken to,
>a data call is a data call....regardless of the speed that it is

Flat-out lie.  Push back and they go back.  A *data bearer capability* call
(that is, one which has its Q.931 Bearer Capability information element
showing "unrstricted digital information" as the b.c.) is billed at the data
rate, even if you rate-adapt it down (V.110 tells you how) to 300 bps.  Data
bearer used to mostly work at only 56k; 64k is now common but not universal
(at least not with inter-LATA carriers).   A *voice bearer capability* call
(that is, one which has its Q.931 Bearer Capability information element
showing "3.1 kHz audio" or "speech" as the b.c.) is billed at the voice rate,
even if you run a modem or DOV across it.

We have several thousand modems' worth of ISDN (PRI) lines provisioned to
carry (mostly) AOL calls.  The calls arrive from the CO as "3.1 kHz audio",
and go into Ascend k56 modems.  The callers are predominantly residential
analog lines.  They are covered by their regular voice plans.

Now let's examine the possibilities.  You call AOL a lot and get crappy 24
kbps response over your noisy paper-insulated local loop, so you take
advantage of the "free" upgrade and pay $8/month to make your line ISDN BRI.
You get a TA with an analog port and plug in your modem and boom, 33.6!
Should you pay more?  Now, you upgrade your 1-year-old 33.6 modem to have k56
capability.  Your analog local loop is 7 feet long (silver satin cord!).
Boom, 53.3 kbps!  Should you pay more?  Now, you decide you don't want AOL
after all and want to call Joe and Andy's Local ISP for $12/month.  They
have a PortMaster, and you get 53.3.  Should you pay more?  Now, you find out
that Joe and Andy's ISP supports ISDN calling, so you figure out how to hook
up COM1 to the TA instead of the modem, and set the TA for 56k DOV.  Since
the PortMaster can thoretically distinguish DOV from "analog" modem calls on
an in-band basis, your call origination is IDENTICAL; only the digital bit
pattern on your B channel changes, from the k56 to the DOV (V.110/56k)
mapping.  Boom, 56k bidirectional!  Should you pay more?

Bell Titanic cannot explain that away.

When you set the BC IE to "unrestricted digital information" (data), the call
goes out on dedicated "Switchway" trunks.  These ALL (so far as they've told
me) go through the tandem, except of course for intra-switch calls, so a
(say) Back Bay to Bowdoin Square call, within Boston Central, goes via
Cambridge Tandem on these "data" trunks.  When you set it to "3.1 kHz", the
call goes out on the regular voice trunks, which only use the tandem when
direct circuits aren't available.  This is arguably a reason why Switchway
costs them a little more than telephone.  DOV, however, follows the phone

>Apparently they have filters on the line to detect modem calls--which
>are billed differently than voice calls, as you have noted. If you use
>56K modem--or any other modem--on an ISDN channel, you will
>be billed as if it were an ISDN data call...according to NYNEX.

Bull crap.  Modem calls INCLUDING DOV (V.110 modem!) are voice.  They do NOT
have filters.  "Detectors" would constitute an illegal wiretap!  They don't
have 'em.  They now work for Bell Atlantic, and MUST tell lies as a condition
of continued employment.  Doesn't everyone understand this by now?  You think
I'm joking?

>Also, beware of anything that NYNEX tells you regarding ISDN charges.
>I have had charges confirmed by THREE different NYNEX employees--
>with time and date of our conversations recorded--and then received
>bills that differ from what they told me. When I tried to get the bill
>corrected, I was told that it did not matter. The moral of the story is
>you don't CARE what your bill is.

I do have a copy of the tariff and I read it while following this thread just
a minute ago.  Of course it doesn't talk about DOV or stuff like that, but
the intent is quite clear, and I *was* a participant in the original rate
case (1991).  Their current "value of service" argument is total hooey,
legally; it's just covering up for a billing error which, rumor has it, is
due to be fixed in a few weeks.

Fred R. Goldstein   k1io    fgoldstein"at"
GTE Internetworking - BBN Technologies, Cambridge MA USA  +1 617 873 3850
Opinions are mine alone; sharing requires permission.

Newsgroups: comp.dcom.modems,comp.dcom.isdn
From: (Floyd Davidson)
Subject: Re: YIPPPPPEEEE! Analog Multilink! (Or: ISDN, who needs it!?)
Date: Thu, 02 Apr 1998 08:11:10 GMT

David Richards <> wrote:
>bill davidsen <> wrote:
>>Floyd Davidson <> wrote:
>>| We agree on that, except I think the price will continue
>>| downward, just maybe not as fast.  ISDN would have been
>>| wonderful if it had been implemented everywhere 10 years ago,
>>| and at rates that amounted to 2x whatever one pays for a normal
>>| POTS line.  The stupidity of telephone companies is beyond measure!
>Then I guess Ameritech isn't as stupid as the average telco, since ISDN
>(both residential and business) does work out to 2x the cost of POTS,
>and is available everywhere in Chicago, and most suburbs.
>Wow- Ameritech did something right?

Not likely! :-(

If they had implemented ISDN in that manner a decade ago, that
would have been something right.  They missed the boat though.

Telco's in the late 80's and early 90's said that ISDN was too
expensive to implement because there was no significant market.
Two things happened as a result.

One is that several modem companies fought tooth and nail over
the millions and millions of dollars in that market which did
not exist.  The v.32/v.34 modem market would hardly have existed
if the telephone industry had implemented ISDN at comparatable
rates (2x a POTS line).  The telephone industry would have
collected all that revenue...

The second of the two items is a little less obvious.  For the
telephone industry to realize where the market actually was,
they would also have had to know then, instead of waiting until
now, that they had a virtual corner on selling access to the
Internet.  The telephone industry owns the infrastructure, and
could have been virtually alone in that business right from the
start.  As it happened, they didn't even know it existed!

If they had been a little more aware, it would have
significantly changed the way that Internet access for the public
developed.  Look at Bill's comments here, and think about what
it would have meant if this type of access had been ubiquitous
at the start:

>>The one thing ISDN buys you is setup time. For some calls you can
>>make the entire call and drop the line in less time than the
>>negotiation of modems with V.34 (and V.8 for setup as I recall).
>>Actually doing this is based on the cost "per dial" vs. "per second"
>>connect charges. But for a demand link between two sites, ISDN feels
>>like a leased line, you don't need to woory about time for the link
>>to come up.

With that type of service in place in the late 80's and early
90's, the requirements for Internet bandwidth would have
skyrocketted even earlier and even higher than they did.  Today
the bottleneck for most of us is the modem connection to an ISP.
Because of the long setup time we tend to stay connected and
cause both the telco and the ISP to have long average call
durations.  For both that means less efficient use of the
equipment (a higher equipment to paying customer ratio).  It
also means more customers are connected at any given time than
are actually sending data because they don't disconnect when it
will be a few minutes before more data is to be sent.  That
means the ratio of clients on line to the bandwidth of the ISP's
Internet connection is higher.  (Which might actually be to the
ISP's benefit, but I work in the telephone industry, so _my_ job
security is not helped by that!)

If ISDN had been ubiquitous by say even 1992, the typical ISP
would probably always have been a major telephone company, data
rates for customer use would on the average be several times
higher than they are now, the connections from an ISP to the
Internet backbone would have to be at least 4-5 times faster
than it is now, and the backbone itself would need be at least
that much faster too.  All of those characteristics mean
customers very happily getting better service and spending more
money with the telephone industry.

In addition, since the telcos generally would make more money
with many many shorter calls rather than fewer longer calls, the
current ongoing perennial flap about long call durations and
hold times would instead be reports of fatter revenues.  And
cable TV companies wouldn't be nearly as likely to spend money
trying to compete for the Internet market either...

It all adds up to the telephone industry missed a big one, and
missed it by a long ways too.  For those of us that have
suffered through the layoffs and other abject stupidities caused
by ignorance in the board room and dimwitted management that can't
manuever forward because they are looking in the rear view mirror
making sure the boat leaves a straight wake, there is little
left to do except send Scott Adams material for Dilbert
cartoons.  It is well not to forget that Scott Adams labored
long and hard during that period developing ISDN applications
for a telephone company, and what he got out of it was Dilbert.
What we didn't get out of it was a better Internet.

Hmmm... after re-reading that I realized that another possible
benefit of a different history...  AOL would not exist.


Floyd L. Davidson                      
Ukpeagvik (Barrow, Alaska)                  or:

From: (Macy Hallock)
Subject: Re: what is a "dml" ?
Date: 18 Apr 1999 11:07:34 -0400

As quoted from wdg@[]:

> In article <7fbn52$116$> Roger Marquis
> <> writes:
> >Frank Heisler <> wrote:
> >>you may very well have a line that will be less than optimal
> >>for modem communications - moreso at the V.90/X2/KFlex
> >>flavoured protocols.
> I get a little irritated when I see ISPs and endusers point the finger of
> "blame" at the telco because their lines and loop electronics do not meet
> the extended bandwidth requirements of V.34 and V.90 modem protocols.

First, there's no doubt that the present modem technologies squeeze the
most out a telco's loop, to the very limit of its abilities.

Second, manys user buy the cheapest modem they can get and expect decent
performance. Not even the best modem's you can buy will make a mediocre
loop or a poor subsciber carrier work much better.  But a good modem helps.

Third, much of this has been brought on by the telco's foot dragging and
obstructionist tarriffs on ISDN over the past 10 years..  ISDN overcomes
all of these issues very nicely and really does not cost much more than
POTS. In fact, ISDN BRI delivers into the premises using one cable pair
the equivalent of two POTS lines, in effect saving cable pairs for the telco.

But the telco's insist the subcriber ordered ISDN must cost more than
2 pots lines...and usually insists on measured service where POTS can
often get flat rated service).  They have the mentality of charging for
ISDN as a premium service instead of as a compliment to basic service.
Of course, the same telco's don't hesitate to use ISDN technology for
"pair gain" POTS lines at POTS prices when it suits them.

Oddly enough, when a telco does make an error and actually price ISDN
in a manner similar to two POTS lines, they get a dramatic increase in
ISDN demand. Then they whine to the PUC about the costs, conveniently
forgeting to mention the cables pairs freed up by this service.
Residential ISDN in Ameritech Ohio's metro areas is an excellent example
of this, but they still insist on "loop repeater" costs on lines
over 18,000', which is more nonsense.

In the meantime, telco's are running promotions to get consumers to
install a second POTS lines in the residence, creating more demand
for pairs and leading to shortages and then to more costly SLC's,
Remotes and subscriber carrier.  Of course, they wouldn't be doing
this if they were really loosing money on each residential line,
like they keep telling the PUC, right?

IMHO, the telco's bear most of the blame of killing off ISDN in this
country...and now they are whining about DSL and CLEC/ISP's bypassing them.
While at the same time they are actively finding ways to deny DSL
access to both ISP's and CLEC's in order to create their own monopoly
in internet access to the residence.

Not that the cable companies are much better, though.
Macy M. Hallock, Jr. N8OBG  +  fax +
  APK Net, Ltd.   1621 Euclid Ave.   Suite 1230    Cleveland, OH 44115 USA

Index Home About Blog