> Were they keeping stats on the line? Are all their error counters showing
> zero? I don't think so. The only thing that might save them (assuming
> its misconfigured) is that T1 circuits are designed to operate at 1.5MHz
> and even with no clock they will migrate to a freq. close to this. So
> with a constant data stream helping the circuit freewheel at close to
> 1.5MHz, their error rates may have been low enough that they never
> noticed. And I guess they just got lucky. But they can't have had zero
> error counts! Science just won't allow it!
>
I believe the above to be what we had, we had a remote pop that we ran
this way for a month or better, we hardly ever maintained over a 2 hour
established connection, but no one ever complained. As long as their was
traffic which was probably most of the time, we maintained an established
connection, but if there was a lag for a few seconds it would drop, then
reestablish as soon as anyone tried to send data. We changed the timing on
our end and we haven't had a loss since, unless we were responsible for
it. And we did have quite a few errors, but since it didn't seem to be
affecting anything, we assumed it to be normal.
We built our little simulated T1 in house with a crossover cable and it
has been up for 6 days, it has actually been up longer, but we had to
power down for something or we would have been up for 2 weeks or better.
Later
Kelley