That includes calls that never made it (e.g. call your pool with a handset,
and it will log that). Also, you dont have the same amount of information
as you do on the PM3s. The ~10% figure is based on what gets recorded in
RADIUS e.g.
LE-Terminate-Detail = "User Request - PPP Term Req"
vs.
LE-Terminate-Detail = "User Request - Call Circuit Closed"
Or for a PM3 today,
normal 91.32% 8167/8943
"User Request - Call Circuit Closed" 7.27% 650/8943
"User Error - PPP NCP Active to Request" 0.84% 75/8943
"Lost Carrier" 0.55% 49/8943
"Port Error - Exceeded LAPM retransmission limit" 0.01% 1/8943
"User Error - LAPM negotiation timeout" 0.01% 1/8943
somewhere between 0-8% is the PM3's fault... Generally, for the people who
do call and complain about being dumped in the middle of a session, its
"Call Circut Closed" thats the problem.
>> Also, a
>> sample of 255 is relativly small. My stats are based on over 100,000
>> calls across a few weeks.
>
>My stats were based on 85 days on 3 busy PM2s.
>
>Measure each day's percentages for 3 PMs for 85 days and you have
>3x85=255 samples. 85 days is 12 weeks and change.
Again, look at the RADIUS termination details. If its based on syslog, you
have the problem of not being able to distinguish between calls that dont
make it to the authentication stage even. The syslog info is not useless,
it just tells you something different.
---Mike
**********************************************************************
Mike Tancsa, Network Admin * mike@sentex.net
Sentex Communications Corp, * http://www.sentex.net/mike
Cambridge, Ontario * 01.519.651.3400
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/>