home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!spool.mu.edu!telecom-request
- From: bcapps@atlastele.com (Brent Capps)
- Newsgroups: comp.dcom.telecom
- Subject: Re: All Circuits Are Busy Now ...
- Message-ID: <telecom13.15.1@eecs.nwu.edu>
- Date: 7 Jan 93 20:03:49 GMT
- Sender: Telecom@eecs.nwu.edu
- Organization: Atlas Telecom Inc.
- Lines: 84
- Approved: Telecom@eecs.nwu.edu
- X-Submissions-To: telecom@eecs.nwu.edu
- X-Administrivia-To: telecom-request@eecs.nwu.edu
- X-Telecom-Digest: Volume 13, Issue 15, Message 1 of 10
-
- In article <telecom12.925.1@eecs.nwu.edu> andys@internet.sbi.com (Andy
- Sherman) writes:
-
- > On 25 Dec 92 21:26:00 GMT, john@zygot.ati.com (John Higdon) said:
-
- >> I would have thought that by now AT&T would have stopped its annoying
- >> practice of drastically reducing its capacity on holidays. A number of
- >> AT&T employees have told me that for reasons that are not very clear,
- >> the company has traditionally blocked off a major amount of the
- >> system's capacity on various holidays such as Christmas and Mother's
- >> Day. This is the real reason you get the "All Circuits Busy"
- >> recording, not because there is an inordinate amount of traffic.
-
- > I can't imagine why they would deliberately turn away business, since
- > they make money selling it. Furthermore, if you go find back issues
- > of the {AT&T Technical Journal} in a library, I suspect you will find
- > that Mothers' Day is the acid test for new routing algorithms like
- > DNHR (Dynamic Non-Hierarchical Routing) and RTNR (Real Time
- > Non-Hierarchical Routing).
-
- I strongly suspect that AT&T turned on every traffic throttling tool
- in their arsenal. That's why they're there -- to keep the switch from
- being overwhelmed by the number of origination attempts. Remember,
- switches have limited internal resources -- call control buffers,
- interprocess communications mechanisms, timing-critical events -- that
- start breaking down when the switch gets really, really busy. If,
- say, a given interprocess communications channel gets choked with
- messages, internal watchdog processes may conclude that part of the
- system is stuck and deliberately swap the switch to the standby CPU,
- which of course will quickly get overwhelmed and lock up. So to
- prevent this they turn on a throttling feature like LLC (line load
- control) which will only give you dial tone on, say, 1 out of every 10
- origination attempts.
-
- > On 27 Dec 92 09:24:00 GMT, john@zygot.ati.com (John Higdon) said:
-
- >> How soon we forget. Hours after the 1989 Loma Prieta earthquake, I
- >> tried at some length to get though to my home from southern
- >> California. All circuits were busy. Then I tried Sprint. The call
- >> went right through. Discussions right here on the Digest pointed to
- >> the policy of AT&T of purposely restricting incoming access to a
- >> disaster area. I, for one, was very grateful for the fact that AT&T's
- >> policies are not always imitated by the competitors.
-
- > The choke after the earthquake had a reason behind it. It was to
- > reserve some large fraction of trunk capacity for outgoing calls from
- > the disaster area. That policy and the reason for it was plastered
- > all over the media, in hopes that people would wait for the "I'm OK"
- > call rather than flooding the network with call attempts to empty
- > houses. I believe that the disaster assistance agencies also liked
- > that arrangement, since it improved the chances of their folks on the
- > scene being able to call out.
-
- This should hardly have come as a surprise to anyone. Disaster
- planning agencies publicised this fact beforehand, and I made
- arrangements with all my loved ones in the state to call a mutual
- friend in Texas who would relay messages and act as a kind of
- information clearing house. When the '89 quake hit I was in Los Gatos
- not far from the epicenter and was able to communicate with my husband
- in San Francisco because we could both call Dallas and leave messages
- for each other. You couldn't get through to San Francisco from Los
- Gatos for several hours after the quake. All in all I thought AT&T
- and Pac Tel did a fine job. After all, the power wasn't back on in
- Los Gatos until the next afternoon, but the phones were working within
- 15 minutes, and I don't know if they ever actually went down.
-
- As an aside, I was working for T1 mux manufacturer DCA/Cohesive at the
- time, and literally hundreds of these big (6' tall) mux cabinets were
- rolling around the test lab and mfg area like loose cannons.
- Engineers were scrambling like mad to keep from being crushed. We had
- a bunch of $5000 Fireberd T1 testers stacked precariously six high on
- top of one cabinet, and it's a miracle they didn't fall over.
- Ironically our customers were required to bolt their muxes to the
- floor instead of standing them on their castors like ours were, but
- one customer's mux fell *through* the floor, so bolting it down didn't
- do them much good. Fortunately, nobody got hurt at our site.
-
- I was in Oregon working for Kentrox on T1 CSUs within two months.
- That quake was the last straw for living in California.
-
-
- Brent Capps bcapps@agora.rain.com (gay stuff)
- bcapps@atlastele.com (telecom stuff)
-