Date: Sat, 30 May 92 19:10:55 CDT
From: varney@ihlpf.att.com (Alan L Varney)
Subject: Re: Fare War Clogs Phone System; Strange Behavior
Organization: AT&T Network Systems

In article <telecom12.430.9@eecs.nwu.edu> coleman@rocky.CS.UCLA.EDU
(Mike Coleman) writes:

> With the temporary reduction of air fares to historic lows, it seems
> that nearly everyone in America is trying to dial the airlines' 800
> numbers.  I've been trying, and like most people, am getting mostly
> slow and fast busy.

> I have noticed that occasionally, at about the moment I'm expecting a
> busy signal, I get a dial tone instead.  Can anyone explain this
> phenomenon?  Is this the "same" dialtone I get when I life the phone,
> or is it a "different" dialtone?

   In general, fast busy is the Reorder tone applied when some switch
doesn't have enough (or any) "no circuit" announcements ("We're sorry,
all circuits are busy now.  Will you please try your call again
later?").  Some 800 providers don't seem to bother to provide the
announcement, so all you ever get is "fast busy" -- you have to figure
out the problem.  ("Fast busy" can also result from a temporary
failure at a switch or database, overload at a switch or database, and
is normally the first-choice alternate treatment for announcements
that are under-engineered.)

   This same treatment can also come from the originating LATA
switches when no circuit is available to the designated carrier --
which might be the LEC itself.  The "fast busy" is typical overflow
for the "no circuit to long distance company" announcement provided at
most LEC Access Tandems.  On the other hand, because of historical
reasons, "no circuit" treatment at your originating switch is usually
"fast busy"; no announcement is provided.

   Dial tone is typically the alternate treatment provided at the
originating switch if Reorder/fast-busy circuits are all busy.  The
tone should only come from your local switch -- some vendors don't
even provide a dial-tone source at a tandem.  Dial Tone can also be
provided about 10 seconds after a call is "killed" by an IXC because
it doesn't have any appropriate tone/announcement circuits.  Again,
you have to guess at the real cause.

   If dial tone is provided very soon (a few seconds) after end of
dialing, it's probably the local switch hitting "no circuit".  If it's
10-20 seconds after end of dialing, it's either the Access Tandem
running out of it's "no circuit" alternatives, or the 800 provider or
terminating LATA running out of the same alternatives.  Between about
four and ten seconds is unusual, unless SS7 is involved in the
originating LATA (might be true for some cases of 800 calls in the two
big Pac Bell LATAs), or you are calling from a PBX.  In the PBX case,
one cannot predict the results, since PBXs don't follow the LSSGR
(well, maybe on their LEC interfaces).


Al Varney - just my opinion

Newsgroups: comp.dcom.telecom.tech
From: floyd@tanana.polarnet.com (Floyd Davidson)
Subject: Re: to clarify
Date: Fri, 27 Mar 1998 19:10:23 GMT

XKMFDMx <xkmfdmx@aol.com> wrote:
>Bob wrote:
>>Thanks to everybody for their responses. Please allow me to rephrase my
>>question. I don't really need an "off-hook indicator" because it would
>>'beep' or 'flash' while I, or someone else, was on the phone. What I
>>want to make, or buy, is something that will sense when the CO has cut
>>off my dial tone because it has been off-hook too long. When I took the
>>voltage readings I didn't find any difference between when I had dial
>>tone (or the "If you would like to make a call, please hang up and try
>>again." message)and when I didn't because the phone had been off-hook
>>too long.
...
>I believe the reason that the voltage readings are the same is that -48VDC is
>supplied to the line even if the CO cuts the dialtone. AFAIK whether you have
>dialtone or not is directly dependant on whether the dialtone generator at the
>CO is coupled with your line. Once the phone has been off the hook for a
>certain amount of time, the CO senses that and applies a reorder message to the
>line.(that really loud, annoying noise) If the loop stays in the off-hook state
>long enough, the CO senses that and disconnects the dialtone generator, and
>presto! no more dialtone.....

A "reorder" is something called a "treatment".  It is a fast
busy signal.

The loud annoying noise is commonly called a "howler", though
there are other names, and I don't recall what the official
terminology is.

The various messages that a switch can put on a line are called
treatments.  The fast busy, or reorder, is commonly used for
several purposes.  It can be used to indicate that a line is not
in service, though that is usually only done on small switches
that have limited capability (a distinct message for an out of
service line is a useful tool for the telco).  Calls that cannot
be completed due to a lack of facilities (all trunks busy) are
routed to reorder, for example.  And commonly a line that is
left off hook and times out without dialing will first go to a
reorder and then to a howler, and eventually that will be
removed too and the line will go "dead".

However, the loop current must continue to be applied because
that is how the switch senses on/off hook conditions.  Hence
there will be no change in the voltage/current conditions on the
loop from leaving the phone off hook too long.

  Floyd

--
Floyd L. Davidson                                floyd@polarnet.com
Ukpeagvik (Barrow, Alaska)                  or:  floyd@barrow.com

Newsgroups: comp.dcom.telecom.tech
From: floyd@tanana.polarnet.com (Floyd Davidson)
Subject: Re: to clarify
Date: Sat, 28 Mar 1998 02:44:32 GMT

Linda K. Sherman <linsherm@gte.net> wrote:
>Floyd Davidson wrote:
>>
>> A "reorder" is something called a "treatment".  It is a fast
>> busy signal.
>
>My understanding of "treatment" is that it's a more general term than
>that, used by some telcos and switch manufacturers to basically mean
>"something you do with a call". This encompasses not just reorders, but
>also routing ("route treatment").

I didn't mean to indicate that a reorder is the only possible
treatment.  As I mentioned, all telco intercepts that go to
recorded announcements are treatments.  A treatment is an
intercept that call can be routed to for other than "normal"
call completion or routing.

>"Reorder tone" is a dual 480/620 Hz tone with a cycle of 300ms on/200ms
>off. "Reorder" is what it's telling you to do--order service again by
>hanging up and placing another call.

Hmmm...  my docs are 10 years old, but Bellcore TR-TSY-000606
says a reorder is 480/620 with .25/.25 seconds on/off.  That is
exactly twice the speed of a busy signal, and in the good old
days we referred to a busy signal as 60 IPS (interuptions per
second) and a reorder as 120 IPS.

>You can have "reorder announcements" also. "If you wish to make a call,
>please hang up and try again" is one of the most often-used ones. You're
>likely to get this when you've completed a call, the other end hangs up,
>and you don't.

That is not a reorder, though it is a treatment.  It is
generally where a "no dial" or "partial dial" timeout is routed
(the line is offhook, routed to a digit receiver, and times out
because a complete set of dialed digits has not been received in
time).

>> The loud annoying noise is commonly called a "howler", though
>> there are other names, and I don't recall what the official
>> terminology is.
>
>This IS the official terminology AKAIK--I've never heard any other terms
>used. I'd be interested if anyone knows of any.

How about ROH, for Receiver Off Hook.  The technical specs are
that it is made up of four tones, 1400, 2060, 2450, and 2600Hz,
and is 15 dB or more above any other tone, with a pattern of
.1/.1 seconds on/off.  But that isn't the good part!  Belcore
even notes that the level should be as high as can be allowed,
because studies showed that for each dB increase in power there
was a 0.5% increase in customer response, which they then
estimated to mean 1 less customer complaint and 1000 additional
complete calls per year in a 10,000 line office.

(I just don't know we all lived without knowing this.  :-)

>> Calls that cannot
>> be completed due to a lack of facilities (all trunks busy) are
>> routed to reorder, for example.
>
>If it's a long-distance (toll) trunk that's busy, they're supposed to
>play a slightly different "fast busy", with a 200ms on/300ms off cycle.
>This allows the dialing end to have a better idea what went wrong as it
>could then automatically try a different long-distance service. It's
>very hard for most people to hear the difference between 200/300 and
>300/200, however.

That would be relatively new.  My mind set is still stuck in the
60 or 120 IPS days. ;-)

>> And commonly a line that is
>> left off hook and times out without dialing will first go to a
>> reorder and then to a howler, and eventually that will be
>> removed too and the line will go "dead".
>> However, the loop current must continue to be applied because
>> that is how the switch senses on/off hook conditions.
>
>What they often do is they temporarily reconnect it at set intervals.
>This is why just hanging up a dead line doesn't always get dialtone back
>right away. Many COs don't disconnect battery any more though, or will
>wait for a lengthy period before doing so. The practice of disconnecting
>is sort of a holdover from the days when the batteries weren't anywhere
>near as good as they are now.

Well, I don't think it really had much to do with using better
batteries.  The idea would be to reduce the load on the battery,
but that comes at an additional expense of including circuitry
on the line card that allows a cutoff to take place.  I'm not
sure how common that is, but it isn't what I've normally seen
happen.

The time delay between hanging up the phone and service being
restored has to do with the way that switching system software
is designed. A digital switch is basically a real-time fault
tolerant online transaction processing computer.  Because it
must be real time (the instant you go off hook you want to hear
dialtone, and the instant you press a keypad key it has to be
recorded before you take your finger off the key, etc. etc.)
there is a pre-emptive scheduling process that controls
everything else.  It will interupt lower priority tasks to run
higher priority tasks.  And when there are no priority tasks
ready to run, it does administrative maintenance at the lowest
priority level.  One of those jobs is to run around and verify
that each call which is set up has a data block with sane values
in it.  Things like call disconnect indicators that are missed
will be caught by these "audit" processes.

Just about the lowest of low priorities is the audit process
that looks for lines with a permanent signal lockout state which
has changed.  As a result, at 3 AM the line will likely be
restored before you can get your hand off the phone as you hang
it up, and a 10 AM on Monday morning it might take 15 minutes to
get a dialtone back.

  Floyd



--
Floyd L. Davidson                                floyd@polarnet.com
Ukpeagvik (Barrow, Alaska)                  or:  floyd@barrow.com