NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/dcom/isdn/1090

Text File | 1993-01-07 | 8.6 KB | 186 lines

Newsgroups: comp.dcom.isdn
Path: sparky!uunet!spool.mu.edu!uwm.edu!linac!att!cbnewse!cbnewsd!varney
From: [e-mail] (Al Varney)
Subject: Re: end-to-end digital connections on ISDN
Organization: AT&T Network Systems, Lisle, IL
Distribution: na
Date: Thu, 7 Jan 1993 06:01:05 GMT
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]>
Sender: Al Varney <[e-mail]>
Lines: 173

In article <[e-mail]> [e-mail] (david.g.lewis) writes:
>In article <[e-mail]> [e-mail] (Joe Picone) writes:
>>We currently have a T1 line installed in our lab that has been
>>guaranteed by the long distance carrier (MCI) to provide fully digital
>>connections once the signal has been converted from analog to digital
>>at the nearest (digital) local office.
>
>I don't wish to denegrate the competition, but I believe your carrier is,
>shall we say, overstating their case.
>
>I don't doubt that they can guarantee that the connection is fully digital
>from your lab, to the serving MCI switch, through the MCI network, to the
>far-end MCI switch, and even as far as the facilities from the far-end MCI
>switch to the LEC Point of Termination/Point of Interface.  However, it is
>physically impossible for an IXC to guarantee *anything* with regard to
>whether connections beyond the POI are digital or analog or how many D/A-A/D
>conversions a signal will go through.
>
>For instance, the IXC may be able to guarantee that all access facilities
>(from the LEC network to the IXC network) are digital.  However, they can
>not guarantee that all Access Tandems are digital.  The connection could go
>from an MCI DMS-250 (digital) over digital facilities to a 1AESS(tm) switch
>AT (analog SPC) (1 D/A-A/D), then over analog facilities to a DMS-100 EO
>(digital) (another D/A-A/D) serving an analog loop (a third D/A-A/D).

   Three minor points:

   1) In the case of a 1A ESS(tm) AT switch, the AT-to-EO connection to a
digital EO (say, a 5ESS(rg) switch or a DMS-100 EO) is just as likely to
be digital -- but that just moves the second D/A-A/D from the EO back into
the 1A ESS switch.

   2) From Joe's later post (below), it's clear that the calls of concern
are moving in the opposite direction -- analog-to-EO-(possibly AT) to MCI.
This doesn't change the number of D/A points possible, but makes some
difference in the number of places the calls can encounter analog.  If
Joe gave MCI a list of TERMINATING numbers, the "digital all the way
to the digital local office" claim could be verified from Bellcore's
NPA-NXX database.  But with an 800 number for ORIGINATING analog calls,
MCI couldn't possibly assure that only one (local switch) conversion
would occur.

   3) Even if the EO (local switch) path through MCI to Joe's T1 line
could be assured, it's possible to run into D/A conversion on the line-side
facilities.  For examples:  You have SLC-96(tm) DLCs that can do A/D near
the analog phone, send digital to the EO and (if it's an analog one) have
a D/A conversion occur before switching and reconversion to digital before
it reaches MCI.  Or the caller may be on a digital PBX, sending digital to
an analog EO (same reconversion needed).

>Unless MCI is claiming that they have direct access trunks to all digital
>EOs in North America, their claim rests on LEC network topology, which is
>outside their control.  And I feel certain that somewhere in North America
>there is at least one digital EO with trunks to an analog SPC AT with
>connections to MCI...

   Fer'sure, fer'sure.  Even if there are direct access trunks to the
digital EO, virtually ALL such EOs will overflow traffic during busy
periods to the AT -- it's unlikely MCI (or any other IXC) would
over-engineer the EO trunk groups just to avoid the overflow.

>>If there is a chance of a reconversion, can some carriers provide
>>special feeds that are guaranteed to be digital all the way? If so,
>>who?
>
>You could always get a private line... :-)

   Probably not for his application -- incoming calls....

>> Some T1s,
>>I am told, actually get converted to analog when they have to pass
>>through a non-digital office (apparently, there are still some of
>>these out there???).
>
>Lots.  The 1AESS(tm) switch was the most widely deployed switch in the Bell
>System; given that they were still being deployed in the early 80's, it's
>going to be a while before they're all fully depreciated and replaced by
>digital switches.

   Take a look at the Bellcore NPA-NXX switch database -- in very round
numbers, there are about 1000 1A ESS switches, about 300 1 ESS switches,
several hundred 2B ESS switches (all analog) in the former Bell system.
Then there are untold numbers of analog switches in the "independent"
and GTE areas.

>>Some carriers claim to deliver end-to-end
>>digital.
>
>I think the big three IXCs all do; the catch, though, is that's end-to-end
>*within the IXC's network*.  Once the call gets outside the IXC network,
>it's dependent on what the LEC's deployed.

>> Since ISDN is new, and point to point connectivity is
>>probably not yet mature, people are speculating that reconversions can
>>occur. I find this hard to believe.
>
>Don't.  There's a lot of embedded base out there.

   David (as usual) is correct.  All ISDN will do is assure you that
the EO (and PBX, if any) is digital -- an analog AT could still enter
the picture.  Or some old analog facilities could exist from the EO
to a digital AT.  These two possibilities will probably near zero long
before the analog EOs are replaced.


In article <[e-mail]> [e-mail] (Joe Picone) writes:
>
>Sorry, I forgot to add an important detail. We supply the callers with
>a toll-free 800 number which, I assume, connects them directly to our
>carrier (MCI) at the local office (equal access, right?).
>
>So, in this case, doesn't this guarantee end-to-end digital if all of
>the carrier's access facilities are digital?
>
>How would ISDN compare (still the same as T1)?
>
>Or is it still possible that in reaching an MCI toll-free number from some
>areas, you may go through some tandem encodings?

   You may have tipped the odds in the wrong direction!!

   Over the next year, the "800 number portability" mandate from the FCC
(aka Docket 86-10) will result in a LEC database dip to select the IXC
for a particular 800 number.  In some areas, this will mean that all
800 calls from some EOs will be routed to an AT, where the dip will occur.
If the AT is analog in such an area, EVERY 800 call will encounter the
"reconversion" from a digital EO or from analog EOs using digital facilities
to the AT.

---------

   On a slightly different topic, Joe, what's the problem with an
extra conversion or two.  Since the analog path is very short in most
of the "reconversion" scenarios, the analog path is likely to have
better-than-digital SNR and very low cross-talk.  And the extra loss
due to multiple quantization effects could be pretty low.  Anyone
have any real-world data on these effects?

   Another thing to watch out for is the cumulative effects of the
A/B signaling bits in non-64K clear digital facilities.  Until SS7
connectivity AND the conversion of facilities to B8ZS, the lowest bit
in any given 8-bit sample might be altered by T1.  In fact, how do
your own T1 facilities from MCI signal on/off-hook?  If it's A/B bit
(or robbed-bit) signaling, you aren't getting completely 8-bit PCM --
every sixth sample will have the low bit forced to 0 (I think it's 0??).
ISDN on your end would eliminate that small effect.

   Note that the actual quality of the sound you get will be affected
by all of the following, in roughly order of highest impact first:

  1) the quality of the customer's telephone instrument (not as big
     an issue with ISDN, but vendor's do use different PCM chips and
     mics..)

  2) the quality and type/length of "loop" from customer and EO (almost
     a non- issue with ISDN)

  3) the amount of junk phone equipment, FAX, modems and screwy wiring
     at the customers site (non-issue with ISDN)

  4) quality of EO to IXC and IXC network facilities, including T1
     and digital switch synchronization.  {{{{ :-)

  5) an extra D/A reconversion over a clear, short analog path.


  So ISDN would remove (almost) the 3 highest voice quality issues from
the picture.  It would certainly give more CONSISTENT quality from
different EOs.  Since all the IXCs are digital, it's no longer possible
to get a really high quality 4-wire analog-all-the-way connection....


Al Varney - just MY opinion