>This is my take on the whole thing.. The fact that it works for MOST
>people shows to me that the code obviously works. If the code was bad or
>at fault here, we would ALL be having problems, therefore I can conclude
>this is not the sole cause here.
Or, we all have different levels of what is acceptable performance, or some
people dont even bother to look at how their units are performing. A while
back a few people posted their termination summary results. Pretty well,
the best statistic anyone was seeing was ~ 90% 'normal' termination
e.g.
normal 89.63% 21251/23710
"User Request - Call Circuit Closed" 8.95% 2121/23710
"Lost Carrier" 0.82% 194/23710
"User Error - PPP NCP Active to Request" 0.59% 140/23710
"Service Unavailable - Failed to detect V.42 remote" 0.01% 2/23710
"User Error - LAPM negotiation timeout" 0.00% 1/23710
"Port Error - Exceeded LAPM retransmission limit" 0.00% 1/23710
Some value between 0% and 10% of abonormal disconnects is the PM3's fault.
Are you implying your stats are closer to 100%, or do you accept 10% of
abnormal terminations an acceptable number.
>
>Most of us have a technical background, very related to scientific. Take
>a look at this in that light. You say your compeitor gets clean connects
>on his equipment. Has it occured to you that it is possible his line it
>set up different, or his copper segment is cleaner?
Yes, it has occurred to me. We are off the same CO, same distance, same
geographical clientel. Of the 16 PRIs, I see the same performance across
all my boxes, with perhaps 1-1.5% variation. At my other 2 POPs, which are
on different COs, I see the same performance on 6 different PRIs. On the
outside chance that my 22 PRIs across all my POPs all suffer uniformly from
low signals, or errors on the lines which dont even register (all my alarms
are consistantly clear and violations are all at 0), I think the only way
to see which box will perform better for me is to eval the unit and see
what happens.
Of the 4 or 5 people who were kind enough to provide me with stats on their
termination causes, the 5300s and the Ascends seemed to be in the 95% range
of normal terminations. The TC unit I was able to study showed about the
same stats as the PM3 -- 10% abnormal terminations, but this was with modem
code on the box that was a year old.
>You people with E1 timing issues and that, I am not directing this at you,
>that is a known issue, and it does need resolved, but I do not buy the
>whole "Lucent's modem code is the cause of ALL my problems" theory.
I dont recall seeing anyone saying its the cause to all their problems.
However, if your competitor gets 5% less random disconnects, or has to get
fewer people to upgrade their modems in order to connect, than its a
competative advantage they have.
---Mike
Mike Tancsa (mdtancsa@sentex.net)
Sentex Communications Corp,
Waterloo, Ontario, Canada
-
To unsubscribe, email 'majordomo@livingston.com' with
'unsubscribe portmaster-users' in the body of the message.
Searchable list archive: <URL:http://www.livingston.com/Tech/archive/>