[aprssig] CQSRVR and JOTA/SCOUTS
bruninga at usna.edu
Sun Oct 21 08:21:21 CDT 2007
> Glad there was so much [CQSRVR] activity yesterday.
> I encourage those of you who want to use APRS
> messaging outside of your area to make use
> of CQSRVR to let people know you are listening.
> Then the actual contact is up to the individuals,
> just like amateur radio is supposed to be.
With this random CQ capability, APRS can make a huge inpact dring Field Day, though of course, the CQSRVR logic will need to be modified to prevent melt-down.
I wonder what kind of moderating mechanism would work to keep the network from total grid lock. My thinking is that since CQ is a "random" process, then randomizing is an acceptible throtteling mechanism too. CQSRVR just sets some kind of throtteling threshold and once the load builds to that amount, then instead of repeating all cq's to all users, that a randomizer just randomly sends it to P% of the total each time. And P is coninuoulsy throttled by CQSRVR to keep the load below the pre-determined threshold.
I have no clue how to estimate what that load limit is. But the point is that we would not want everyone's local RF load to increase by more than say 20%? But then CQSRVR has no idea how many users may be in one area and so some RF areas with lots of users would saturate much earlier than an area somewhere else with fewer users.
A complex method requiring lots of effort, would be for CQSRVR when it is in throtteling mode to quickly check the LOCATION of the sender. And then when it sends things back out, it looks at the LAT/LONG DEGREE digits. And on each re-transmission, it keeps a bin count of where each outgoing is sent, and only allows one copy per bin to be sent. a 5 by 5 grid of lat/long is 300 by 300 miles and whould pretty well isolate RF areas. There are maybe 40 such bins over the USA..
There would still be more throtteling in addition to this, but this helps assure that the random throtteling is more equally distributed and cannot overload certain RF areas.
In fact, the bin counts could also accumulate over time. And say not allowed to send any more than X packets per 10 minutes in each bin. And now this could be an easy way to set the threshold... In any RF area, we would not want to say see more than 1 CQ per minute. In this case, then the throtteling can make sure not to send more than 10 packets per 10 minutes to each bin... something like that.
Besides, everyone in the RF area of a given IGate
will see the outgoing CQ's to other stations in its RF area even if it is not addressed to them, so they will still be able to see new callsigns to work on the air... eventually.
Just somthing to think about before Field Day 2008..
More information about the aprssig