[aprssig] Ultimate APRS digi

[Robert Bruninga]

>> 1) Everything on every alternate input is 
>> digipeated over to 144.39
> 
> This scheme helps but it doesn't scale.   
> I think the 5x to 10x is overly optimistic.  
> it might take the 144.39 channel from very
> overloaded back to just overloaded.  Then a 
> few more users show up and we're back to 
> very overloaded.

It can scale to any size just like the cell phone system.

And it is easy to see the benefit.  
1) Just monitor one's local digipeater.
2) Count only 1 hop packets BY THAT digi
3) Count all other packets your station hears on the channel.

The ratio of those two numbers is the ratio by which local
performance is improved by having mobile and other inputs on an
alternate channel.  In my area it is 10 to 20 to one
improvement. The key here is that people are overlooking all the
congestion on the channel caused by the outputs from ALL the
digis and not realizing that you compare that large number with
only the packets that a SINGLE digi hears direct.  And that
ratio is the improvement made AT EACH digi by having a user
input-only channel.  Input packets can only colllide with other
local-this-digi-only packets.  And there are much much fewer of
them.

Of course, once the performance for all local users improves so
much that each digi is starting to be more than say 80% key down
(With everyone getting in very reliably), then to scale further,
we simply have to reduce the input range of each digi and add
more of them.

> Expecting to see all traffic locally in a busy 
> area is probably asking more than a collision 
> hindered 1200 baud channel can provide.

It is the old ALOHA concept.  APRS Clients that have implemented
the ALOHA circle calculator can show what the size of a full
channel is.  In Wyoming, that circle range might be 150 miles.
In LA, it might be 15 miles.  That is a 100-to-one area scaling.
Any area that gets down smaller, will just have to have
telephone-pole digis every 10 miles.   Once the ALOHA circle
gets down to 5 miles in most areas, then there is no need for a
digi.

I agree, that when we get to more than about 100 users in 5 mile
range of each other that we can go no further at 1200 baud.  But
that is why this plan includes 9600 baud at the same time.  It
lets us transition with growth while still supporting everyone
with legacy systems at 1200 baud.

> Using an 9k6 flood of all traffic in the 
> area and multiple 1200 input
> channels may be more workable.

Yes, that is the true ultimate.  And areas that are putting up
the parallel 9600 baud systems will be the first to get there.
Thank heavens that all the APRS radios can already do 9600 baud
easily.

Bob, WB4APR