[aprssig] APRS low-power-local ALT input channel

[Robert Bruninga]

>>> HamLists at ametx.com 9/25/04 4:05:08 PM >>>
>As pointed out already, we are not running an ALOHA 
>network, but a combination ALOHA/CSMA network.  

Yes, but the network is 99% ALOHA to the users on the
ground and only about 1% CSMA.  Anyone can satisfy
this to himself by simply looking at how many users he
hears per hour on 144.39 versus how many users he hears
*direct*.   Most users will find the result about 1%.

Thus your position that the network is CSMA and that all
operations *MUST* operate based on the CSMA rules are
just trying to apply an academic rule to a real situation that
simply does not exist.

>What you are advocating is something that the University 
>of Hawaii determined to be untenable: a pure ALOHA
>network with multiple stations on the same channel.

The Univ of Hawaii found NO SUCH THING.  In fact their
studies are the basis of all ALOHA networks of today, and
hence the name "aloha".  What they found was that the
maximum throughput on an ALOHA channel peaked at
about 18% of possible channel capacity and that adding 
more stations caused a precipitious drop in reliability and
throughtput above that point. 

They then went on to show that if you make the links full
duplex so that all stations CAN HEAR ALL OTHER STATIONS
then they can use CSMA to almost double channel capacity
to near 36% or so.    But only the digipeaters in APRS can
hear each other.  Hence they can bring in EXTERNAL 
UNWANTED QRM packets more reliably, but this does
absolutely nothing to improve the reliability of the local user
to be heard.  He is still operating in a 99% ALOHA network
with his fellow users.

>Bob, I don't where you cook up your numbers, but 
>CSMA in the Dallas area (over 100 users) is closer to 
>80% effective.

Sure for anyone running 50Watts from a good base 
station antenna.  In fact, he can be heard even if the
channel is 100% overloaded, because he simply
clobbers everyone else.   Your use of a term like
80% effective can mean that the otehr 20% never
can be heard at all...

Lets see your numbers.  Look at any users USER list.
Compare the TOTAL users heard, versus the total
heard direct.   That is the ratio of stations operating 
in the ALOHA regiem.  And I am not including DIGI's
heard direct. 

Your issue seems to be with coliding with other users,
and those are the only ones you can avoid a colllision
with if you can hear them DIRECT.  SO tell us how
many stations your MOBILE can hear direct versus
how many hundreds of stations are on the air per hour
in Dallas and then we can have apples and apples
to talk about.

In my area with a 6 dB commercial 22' stick at 80 feet
above average terrain in the Balto/Washington DC area
I hear about 300 stations per hour and only about 6
direct.   Oh, ok, make it 2%.  Sorry.     THe 1% is because
half of those 6 were mobile and just happened to pass
within a few miles of my house.  When I came up with
the 1% figure, I was talking about permanent always
direct stations...

> And what about everybody else who says "I'll go over 
>to that channel so I don't have to compete for 144.39".  
>The flaw in your design (among others) is the fact that 
>you have no way to police the frequency and prevent 
>everyone from moving over to that frequency.

Ah, if every user and mobile moved over to the alt freuqncy
it would be the greatest thing for APRS that has ever happened!
That is the *whole point!*  As I said before:

A DIGI hears:
1)  a few locals 
2) and hundreds of out-of-area packets.

An alt digi only hears 
1) a few locals!  

This separation of the INPUTchannel from the OUTPUT 
channel is the first thing any ALOHA system can do to provide 
an orders of magnitude improvement in local reliability by not
making every packet from the user have to compete with 
the other 98% of packets that are not local...

>Wrong!  I fully understand how UI AX.25 packet works.  
>And 99% of APRS depends on CSMA, not transmitting in 
>the blind.

I am not advocating transmitting in the blind!  I have no
idea where you are getting this stuff.  I am saying that
APRS users who listen to the channel before transmitting
only have a very-very tiny probability of hearing any other
user.  So your arguments against having an alternate 
channel listen-only input to a repeater just make no sense
in reality...

de WB4APR, Bob