[aprssig] The new N-N paradigm summary (Cap)

[Robert Bruninga]

>WIDEn-N APRS network is amazingly resilient!  
>Warnings of a VHF net "death spiral" have been 
>greatly exaggerated.

I too can't resist commenting on the above statement.

And that is the problem!  WIDEn-N will ALWAYS work.
And it will ALWAYS help your system to see MORE stations.
The problem is, the more you see, the less the chance that
a local tracker or portable can get in.

The 50W station might get in well enough.  Or the 1 Watt station
living near a digi might ALWAYS get in.  But both of these
see *good relliability*, but at the expense of someone else that
they collide with... always.

I think we have to hammer home to all users, that if you
are seeing much more than about 50-60 users on RF,
then by definition, that means the less reliable for locals
your network is.  The success of APRS on RF is not
how many you see, but how reliably you can communicate
locally (nearest 60 or so).

Bob

original message from dick:
*------------------------------------
Resilient also means that when the local network gets saturated, nothing
happens. I live on the fringe's of 3 major metro areas the nearest being
about 145 miles away (LA, San Diego and Phoenix). My local station was
hearing over 180 stations on RF. Got to the point we couldn't hear our own
local stations direct.

We dropped WIDEn-N and TRACEn-N at 2 fringe digipeaters and local stations
heard dropped to 50-60 instantly. For the first time in years, we actually
could see our own mobiles using WIDE only. The local network now works.

Your network might be different and that's a blessing. The west coast is
saturated to the point that not all packets are seen. Common sense is NOT
setting your mobile to WIDE5-5 because you think your packets aren't being
heard by 10 local digis. If mobiles simply set to WIDE only or RELAY,WIDE
the packet success rate would increase.

>More of our users are learning that
>simply WIDE2-2 is a good path for general use, and it works.  It's better
>than a path that includes any commas.  Using only WIDEn-N or TRACEn-N
>produces no dupes.  Kantronics TNCS do not do any dupe checking on packets
>other than with pure WIDEn-N and TRACEn-N paths.  RELAY or WIDE or TRACE
>aren't much needed, and digipath commas cause loss of dupe-checking in our
>Kantronics TNCs. Going back to RELAY,WIDE (or worse yet RELAY,WIDE,WIDE)
>would be a step backward in California.

WIDEn-N and TRACEn-N is like adding training wheels to the local network. It
is used because education failed. That's like saying 5 watts is ok, 50 watts
is better in downtown LA. Crank it up to be heard! One of the 10 digis will hear
it. Pretty sad.

Several years ago LOAD monitors were installed in the 3 major locations. They
worked fine until they heard 250 packets in 10 mins. Then, they just couldn't
hear anymore because of collisions and saturation.

Yup, I disagree with Caps and Earl's generalizing. They both are fortunate at
the moment. I guess they don't see those stations 500-800 miles away flooding
into their own networks... the problem is that we saw 'their' networks in our
area <g>.

A step backwards sounds like a good plan. RELAY,WIDE is much better than WIDE2-2
that propagates packets in all directions because we don't understand what
effect our local digis have on adjacent networks. More Igates and better designed
digis cover the distance just fine.

Dick, KB7ZVA
APRSWest  

 



_______________________________________________
aprssig mailing list
aprssig at lists.tapr.org 
https://lists.tapr.org/cgi-bin/mailman/listinfo/aprssig