DH>> Since then, some careful politicking has turned up an
DH>> admission from a senior Telstra manager that there _is_
DH>> dynamic line bandwidth sharing (read limiting) hardware
DH>> now fitted on high traffic trunks and exchanges.

TM> Hmmmmm, very interesting. This would explain the problem I've had
TM> here in Tullamarine keeping connected for reasonable periods of
TM> time at 28,800 (26,400 is better, 19,200 is fine).

TM> Would any Victorians reading this know whether Tullamarine is
TM> classed as a "high traffic" area?

Observation seems to be that it's more a time of day thing than a
"route" thing.  (This might not be true on a high density high
value trunk, granted that such a thing still exists in this day of fiber
everywhere.)

My miseries concentrate on the 2 PM - 9 PM window, for reasons of their own
- the utter worst concentrating around 5 PM - 7 PM.  This makes very little
sense to me - it doesn't match a traffic pattern, other than the famous fax
back to head office in Japan.

It doesn't correlate to offpeak power control tone, which is the first
thing I thought of when I found the pattern.  It _could_ match at the
exchange, granted that the 2k distance is enough to be in a different
service area.

The dramatic "disconnect now" effect, on one direction only at a
time is characteristic - and it only happens when a high speed modem is
present.  If it's 9600 or less, the connection never sees a problem.

If the problem _is_ triggered, the SNR is less than 10 db.  If it isn't -
the SNR is over 38 db to almost all my callers.  Frustration is.

Cancel of the worst line also is.  Next week.

Regards,
Dave Hatch.

--- Msgedsq 3.20