Network - tweaking and perfecting

[Chad Kurszewski WE9V]

>I received a disturbing report last night that the network "sucks".

I wouldn't go that far.  It's 99% great, needing a slight improvement.

>The early indication is that the 30 second retry was too long.  Probably
>this station was using it for a multi-single and if a QSO didn't get 
>updated immediately, they had to wait 30 seconds for a retry.
...
>Maybe there was an RF problem...


I operated SO2R with 2 computers for SS SSB, ~1900Qs.  I would agree that
the 30 second retry is too long.  I'm positive that the temporarily missed
network traffic was not due to RF, as I seen it while not transmitting,
as well as never experiencing any RF problems in the shack.  (The network
cables are run right alongside all six 1.5KW coaxes, 4kV feeds, etc, for
years and never a problem M/M, M/S or SO.)

I was using 2 P5-130 machines.

I didn't use the 2 radio method at the higher rates, as this would have
definitly caused some serial number discrepancies.  I did find myself
waiting for the network to catch up and, when tired enough, I handed out
one busted QSO #.


I think there are two problems here:

1) Why isn't the traffic making it the first try?
2) 30 Seconds is WAY too long for the first couple retries.  I think the
   first few retries must be done as quickly as possible.  There must be
   something in the queue code that even if the message gets resent a
   few times before being returned, that it does get cleared from the queue
   when returned.  After the first few (2-3?) quick retries, then retry
   at a slower rate.

---
Chad Kurszewski, WE9V              e-mail:  Chad_Kurszewski@csg.mot.com
The Official "Sultans of Shwing" Web Site:  http://www.QTH.com/sos