[aprssig] back to tactical aprs

Andre v Schaijk aprs at pe1rdw.demon.nl
Fri Dec 10 05:57:01 CST 2010

Bob has been sharing his frustration that aprs has become unusable for 
tactical communication thanks to the abundance of mobile trackers.

Offcourse splitting tactical and tracker trafic (and combining them on 
uhf) is a good fist step but over the years another problem has made 
it's way into the network.

It starts with using level 2 routing where it should be handled at level 
3 and set at level 7, for those not familair with iso network layers 
level 1 is the hardware (radio/audio/1k2/9k6/etc) level 2 the transport 
protocoll (ax25) level 3 is routing (net/rom flexnet inp3 ip etc) level 
4 transport (tcp aprsobjectname etc) level 5 session (end to end 
connection etc) level 6 presentation (aprs data text etc) level 7 
aplication (win aprs, ui-view etc).

right now we are setting the number of hops you want to go at level 2 
(wide2-2) but for tactical use we should not have to worry about the 
network, all we care about is how far it gets, this should be a setting 
in level 7 that is forwarded to level 4 and used by level 3 (the digi 
checks how far it is from the sender and decide if it has gone far enough)
On a transport level we are handeling duplicate reduction of level 2 by 
abusing a level 1 setting (setting dwait to 0 making sure we get on air 
collisions, this is against normal network planning, it make much more 
sense to use the digi software to decide if a packet is a duplicate or 
not (like digined does) this will safe onair time compared to using 
level 1 collisions because FM capture is not perfect and leads to 
undecodable packets and time shift between digis thanks to differences 
in txdelay and difference in reception of hidden nodes.

I'm fully aware that it grew this way because at least in the usa a lot 
of sites depend on embaded hardware to do everything but today it makes 
no sense to stick to outdated hardware (8bit z80 based tncs) where 
embeded pc systems have simular power consumption and can do a lot more.

73 de Andre PE1RDW

More information about the aprssig mailing list