[aprssig] Why do routes go very wrong?

[Stephen H. Smith]

  On 8/12/2010 10:19 PM, Jaye Inabnit ke6sls wrote:
> Hello all:
>
> My I-gate is ke6sls-3, my remote APRS unit is ke6sls-7.
>
>
>
> I conduct several more tests only to find that again, APRS-IS is routing the
> packets to the wrong gate.  Minutes later, one or two packets (out of many)
> are routed to my I-gate and then immediately ack'd yet again.  I'm attaching
> my rf-log so you can see what my server sees.  My server is responding
> correctly to each of the packets received from either APRS-IS or RF.
>
> I'm completely stumped--it's like the northern station has somehow hijacked
> routing--though that is plain silly, but still--why is this errornous routing
> taking place?
>
> Then there is the question of why the station "n6nar" is involved in the digi
> route.  I must be missing something obvious here but I'm sure I don't know
> what that something is.
>
> What is happening here and what can I do to make sure routing is correct?
>


1)     The station after "qAS" is always the call of the igate station that 
took your transmission off RF and inserted it into the Internet system.   N6NAR 
is the igate station.

2)     You seem to be using a two-hop path, meaning your transmissions are not 
only being heard directly by your own igate, but also by digipeaters and other 
igates possibly hundreds of miles away.

3)     Most digipeaters and the APRS-IS consider any additional packets from 
the same sender with identical payloads to be duplicates,  -IF- heard within 30 
seconds or so, and will ignore them.   If a packet is heard MORE than 30 secs 
later, it is normally considered "new" and will be processed.   I.e. 
"last-man-in" wins -IF- the last man is at least 30 secs later.

4)     If a man-in-the-middle digipeater is delaying retransmitting packets 
(perhaps it's receiver squelch is being held open by distant activity you can't 
hear on the ground) more than the dupe-detection interval, it could fool a 
distant igate into thinking it is hearing a new packet, rather than a dupe of 
an old one.  In turn, the distant igate and it's APRS-IS port will now think 
they are the one that should handle the reply, and will reverse-gate any replay 
(or ack) to the message long after your home station igate did the same thing.

Try a path with a single hop (i.e. just WIDE2-1) and see what happens.    Or 
even no hop at all for test purposes.