[aprssig] Re: Tracker Smart Pathing: user types, alternatives

[apratt at bestbits.org]

I'm really pleased we've been able to carry on this discussion as well
as we have. I like being able to craft several points in a message and
see cogent responses to most or all of them in the replies.

On a side note, I wasn't clear enough in making one point before. By
"local user" I meant one who lives and roams in a given area, and who
can recover from a digi's bad auto-routing by reconfiguring their
tracker. As for me, the wide-ranging traveler, I'll just suffer an
outage as I pass through. C'est la vie.

If this "auto-pathing" idea requires new digi infrastructure, clearly
it's a non-starter. Ideas that require new digi infrastructure should
set their sights higher, with true network-managed pathing, so a user
can really say "Get me to an IGate" and trust the network to do it.
I'm not convinced, however. I think there's a win here if we could
just get to a few critical digis: those in the cities (trap down to
one) and 3-hop areas (bump up to three). Then I could run WIDE2-2,AUTO
nationwide without replacing *all* digis.

It sounds like we're halfway there, though. I love the idea that the
LA Basin will trap my rude, abusive 3-hop packets down to one hop as I
pass through. That's news to me, and welcome. An earlier description
of trapping (on the New n-N Paradigm page) says digis should DISCARD
packets that specify excessive hops; repeating them just once, and
doing that even for usually-not-abusive 2- and 3-hop packets, is much
better for me.

But it begs a question. One hop in the LA area is just fine for my use
case, but why is it acceptable for yours? The mission statement for
Proportional Pathing is "To be visible over the air both locally and
multiple hops away, with decreasing update frequency over distance in
order to reduce congestion." This user is not served by having all
packets trapped down to one hop. It seems like this would be anathema
to you.

But hey, if you're OK, then I'm OK. If most places with congested air
are willing to trap packets down to one hop, I think we're just about
home.

Did you know that a TinyTrak 3 can have two configurations, and can be
told to alternate between them with every packet? If I set one of the
configs to two hops and the other to three, and set the rate to 3-5
minutes, I think things will be pretty sweet. In an area with
congested air like LA, I'll get one hop for every packet, which is
enough and not rude. In other areas, I'll alternate between two and
three hops. This should give an OK success rate in the wilds, and is
not *too* bad when I'm in a one- or two-hop area that doesn't trap.

This is effectively three-step Proportional Pathing, with two of the
steps implemented in the tracker and one implemented in the big-city
digis. Stick a fork in me, I think I'm done.

-- Allan Pratt, apratt at bestbits.org