varney@cbnewsd.ATT.COM (Al Varney) (11/07/89)
In article <telecom-v09i0482m03@vector.dallas.tx.us> judice@kyoa.enet.dec.com (Louis J. Judice) writes: >I never realized this until it was demonstrated to me, but at least in >NJ on the Peapack Central Office, if you call person "A", person "A" >can hang up, and pick up their phone up to about 15 seconds later >without disconnecting "B". This is without any phone features, etc. >If the calling party hangs up, of course the conversation is over. I >also suspect that if the "called-party" is on a PBX, etc., that this >"grace period" is not given. >What is this, why does it exist? It is "dependable" or just a fluke? >Is the 15 second limit a standard of some sort? See "DISCONNECT" below: And in article <telecom-v09i0488m09@vector.dallas.tx.us>, westmark!dave@uunet. uu.net (Dave Levenson, 1 Nov 89) writes: >This is generally true throughout NJ Bell territory. If the calling >party goes on-hook, the call is disconnected (the called party is left >high-and-dry for about 20 seconds, and is then given dial tone). But >if the called party goes on hook, the call remains up for about 20 >seconds. If the called party goes back off hook during this "grace >period" the conversation may continue. I'm not sure this is >_universally_ true. It is certainly not true where the called party >is behind PBX, unless the PBX implements the same "grace period" >feature. DISCONNECT: change from off-hook to on-hook that persists beyond a prescribed time limit and can last indefinitely thereafter. (Bellcore LSSGR,6.3.5.2) But you want details? OK. "True" disconnect (no Flash allowed): 0-200 ms Hit { "false" disconnect?} 200-400 ms Hit or disconnect {each switch draws the line within this range} > 400 ms TRUE Disconnect Disconnect (Flash allowed - e.g., line has 3-way calling) 0-200 ms Hit 200-300 ms Hit or Flash 300-1100 ms Flash 1100-1550ms Flash or disconnect > 1550ms TRUE Disconnect These apply to caller and called parties. When CALLER disconnects, the CALLED party gets no dial tone for 10-12 seconds because, in most cases, they don't want to make another call. Good switch design tries to give dial tone to the lines most likely to be able to complete a call. Most of the time the CALLED party doesn't disconnect right away, the party either missed the switch-hook or is still scribbling down the message, etc. When CALLED party disconnects first, a 10-12 second "timed-release" period starts. The connection to the CALLER remains, such that a CALLED party off-hook will result in a stable talking connection again. Why? Don't know, but a good guess is historical. The party answering might answer, realize it's for another household/business associate, hang up and yell for them to answer their extension. Alternatively, the answering party could hang up and run to another extension. The reason could also be related to false disconnects when operators dropped off or bridged on to TOLL or COIN calls via switchboards. For old-timers, the term DISCONNECT applied to only the CALLER and the term HANG-UP was reserved for the CALLED party. For inter-office trunks: 0-150 ms Hit {in talking state, shorter in other states} 150-500 ms Hit or disconnect > 500 ms Disconnect The above is a condensation and re-write of information in the Bellcore LSSGR, TR-TSY-000506, a Module of TR-TSY-000064; the AT&T Practice 781-030-100, "Notes on Distance Dialing", Iss. 1, 1975, later called "Notes on the Network" and then re-titled/edited as "Notes on the Intra- LATA Network" by Bellcore after you-know-what. PLEASE don't design any equipment or program based on the above information. BUY Bellcore's documents (and AT&T's, if they apply). I've left out 4 pages of exceptions, extensions and special cases that will allow you to almost work in the real world, but not quite. And finally, in <telecom-v09i0492m06@vector.dallas.tx.us>, lars@salt.acc.com (Lars J Poulsen, 3 Nov 89) reads: >In article <telecom-v09i0486m09@vector.dallas.tx.us> > MJB8949@ritvax.bitnet (Nutsy Fagen) writes: >> In a relatively simple manner, could someone explain how a >>local telco communicates with a PABX in terms of incoming/outgoing >>phone numbers? .... and he (lars) writes: > ANI (Automatic Number Identification) ... On an outgoing call, > the PBX tells the CO which extension the call came from. > ... a phone-company owned "PBX" (I put it in quotes, because being > phone-company-owned, it is not "private") is called a TANDEM switch. In #1/1A ESS(tm), it's called CENTREX-CU [CUstomer premises, vs. the more "normal" CENTREX-CO [Central Office]. It's still mostly a class-5 switch, just on non-TELCO property. One of the major differences between the PBX ANI (used to be called AIOD-Automatic Identified Outward Dialing(?)) and REAL ANI (sent only between Central Offices and other TELCO equipment) is the PBX could be slow or broken or just plain lie! The older AIOD data link basically indicated the low-order digits of the last outgoing call -- too many calls at once and the "last" call and associated digits could be confused. It's major purpose is in generating Call Detail records for the PBX customer: The AMA Billing still goes to the designated Billing Address for the PBX. Most "modern" PBXs generate Call Detail records locally, along with more data than most administrators could possibly use. (This applies to switches also.) > A PABX (Private Automatic Branch eXchange) is a smaller version of > what the phone company has in the central office. .... #small flames Common misconception; a PBX/PABX is not a central office switch stripped- down. PABXs don't follow most of the rules of central offices (like the DISCONNECT stuff above) and almost none of the rules of trunking. Each vendor can do as they please as long as the customer is happy and the TELCO (or by-pass carrier) doesn't object. The rules in the LSSGR, etc. are there for a reason (too bad Bellcore doesn't give the reasons), and many of the reasons are there because of Operator/Emergency Services, TOLL requirements, outside plant facilities, CAMA, AIS and a requirement to be compatible with things that exist in only a few places (and in requirements). If PABXs could operate with minimum customer complaints, accurate Call Records and maximum reliability in the real world of backhoe fade, lightning, open wire facilities, digital facilities that fail and fade Off-hook, cheap phones /answering machines/ modems/other PBXs, EMR, foreign voltages, abusive customers, World Series overloads, two-party lines, 8-party lines, cord boards, COIN lines, earthquakes, heat, humidity, dry heat, cold and non-stop operation, maybe they would approach a central office in capabilities, even if they don't have the size to support 90,000 lines. #flames off Sorry, I realize PABXs are complex and do some things a central office would have a hard time doing (including meeting the National Electric Code), but all those standards for REAL switches aren't in there just to fill space. One MAJOR casualty of divestiture is the knowledge base that knew WHY the specs said a particular thing. Al Varney, AT&T Network Systems, CCS7 Network Services Customer Support, Lisle, IL This note and I are NOT official spokespersons for AT&T