> The irony of a Canadian commenting on
Actually it probably wasn't patent infringing. The 911 message protocols are
in the "public domain" and could be interfaced by anyone (I wrote an interface
for San Antoneo, Texas before I worked for the 911 group). The problem came
when other changed the system.
One of AT&T competitors beat it's response times on the PSAP (Public Safety
Answering Point) by anticipating a call.
AT&T system: with 2 second REQUIRED initial response time:
receive call (with telephone number) -> pickup phone -> send request for
lookup->
display results
Competitors system:
receive call (with telephone number) -> send request for lookup-> pickup
phone ->
display results
Our competitors system was good until the San Francisco earth quake. The 911
system was swamped and many people would hang-up and then call back. Our
competitors system would send every call to be looked up even if they did hang
up. When the backlog was flushed the response time for lookup had reached 30
minutes and for all intents ALI (Automatic Line Identification) was dead. I
hope no one else
died due to these changes.
Our competitor was forced to change their system to conform to post phone
pickup to send their request for lookup.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
> On Tuesday 07 January 2003 23:28, Scott Siebold wrote:
- From a software engineer/parallel software design PoV, the solution is
to use
their process until incoming calls pass a pre-determined rate of
inbound-requests, then switch. May take some tinkering to determine the
exact threshold, and it'll depend on the reference system's ability to deal
with accesses, too, but its doable.
ObGZG: Ok, I'm a geek, I couldn't help it.
- --
Alexander Williams (thantos@telocity.com) | http://www.chancel.org
|
<snip>
> ObGZG: Ok, I'm a geek, I couldn't help it.
Gracias,
AlexW:
> ObGZG: Ok, I'm a geek, I couldn't help it.
Glenn:
> Aren't we all in one format or another?
<grin> my brother has suggested that most of us on this list probably have
some degree of Asperger's Syndrome
(http://www.udel.edu/bkirby/asperger/)
(as do many Trekkies, train watchers, stamp collectors--and physicists,
mathematicians, etc)
> Hash: SHA1
It isn't the answering point that was being overloaded it was the point where
the lookup was being done at. The originating point must be allowed to send
the request for a
lookup as quickly as the calls are answered or the system will not work. The
only connection between answering point and lookup point is a dedicated line
(actually two lines) over which the request for a lookup goes out and the
response comes back.
I don't know the details about California but in the midwest four states
(Illinois, Indiana, Wisconsin and Michigan) are managed out of one location
(Indianapolis). Local disasters happen regularly (plane crashes, train
derailments, fires, etc.) and may cause major increase in demand locally but
do not effect the system. It's when demand increases over a large area and
multiple sites start maxing out that the demand has an impact.
It was AT&T that designed the system and as designed would not be overloaded.
The competitor was looking no further then the local answering points and so
the speed advantage was used as a selling point.. The problem was not
discovered until a large area disaster occurred.
I'm also a "geek" and have no problems with other "geek" comments.
From: "Alexander Williams" <thantos@telocity.com>
> On Tuesday 07 January 2003 23:28, Scott Siebold wrote: