[aprssig] packet stats

Wes Johnston aprs at kd4rdb.com
Tue Dec 21 14:02:43 EST 2004

I'd never heard of AprsNetSpy until last night, so I decided to give it a try.

I ran one copy hooked to kc4pl.zapto.org's javaarps filter.  A 2nd copy is
hooked to a local TNC.

The java filtered port is limited to the area in which we would like to see
coverage... the state of south carolina.  We limited the filter range to show
charleston, florence, columbia and greenville.  We have managed to avoid
charlotte traffic because they would be on the NCn-n system instead of SCn-n.

After running for 12 hours, here are some stats.

                Filtered    On-air
total packets   1173        5509
posit reports   710         3395

in the past 60 minutes as of 7am,
posit's/hour    113.5       531.1

Now 531.1 packets per hours may not seem high, but that is packets _decoded_ per
hour.  The only way I know to measure the packet which aren't decoded would be
to hook an hour meter to the DCD led and assume 1.5 seconds / packet for the
accumulated time.

The point I hope to make here is that the difference between LOCAL packets that
I care about vs DX packets which are clogging our network is about 4.7:1.  If
the lnkN-n and SSn-n systems were fully implemented, we'd see nearly a 500%
drop in network utilization... and my local folks would have a more reliable
network.  In other words, we have a network which is capable of allowing me to
communicate with more stations than I ever could in the PBBS/CONVnode days, but
let's not hammer it 100% all the time.  If we can thin out some of the stations
streaming in, our local networks will once again be really reliable, and if the
traffic stays low, we could actually get messages out further than seems
possible today.


More information about the aprssig mailing list