Fwd: Re: [aprssig] Packet routing, path specification.

[Robert Bruninga]

>>> HamLists at ametx.com 06/23/05 5:53 PM >>>
>Your example of wanting to check into the Philly 
>APRS net on Sunday's.   Why is it so important 
>for you to do this via RF?  

Because I have always enjoyed HAM radio and
using it to do things and doing it while mobile and
other interesting locations.   Nothing wrong with
playing with the internet, but it isnt ham radio.
(though I have nothing against free bandwidth
no matter where it comes from).

>If you send a message (that is what checking into 
>a net is, right?) to the net control, your message 
>and your position will be gated to the Philly LAN.  

Around here we dont see any kind of success with 
that APRS-IS feature at all.  Using that technique, 
the position is sent ONCE and ONLY once.  If it 
collides, it will be a HALF HOUR before the Igate 
will allow another courtesy posit.   By then the NET 
is over.  Humans are gone.  Nobody to play with 
any more.   It simply does not work reliably enough 
to even pretend that it works..

>Gee, no need for discrete paths.  That is the key 
>for getting rid of source routing COMPLETELY!  

Yeah, I konw,  use the internet...  There are 
lots of people also that say who needs HAM radio.
Get a cell phone...

>APRS has _no_ demonstrable need for source 
>routing (at least you have yet to provide one 
>example other than "because I want to").

I have too, and you are absolutely correct, that
the main reason is so that the INITIATOR of
a communications NEED can send the packet 
where HE WANTS it to go.  Only HE konws 
what his immediate need is.

>There is no such thing as "ROUTINE and IMMEDIATE" 
>requirements for getting your non-message 
>packets to a distant area on the user's say-so.

I disagree.  THere is a world of difference between 
home stations sending 24/7 posits doing absolutely
nothing but routine keeping the lights on, compared
to those hams out there doing something such as
WX spotting, or supporting a local marathon or bike 
event, or doing a local club net....

>Bottom line: As I stated in the beginning, those 
>that will feel the "pain" of NSR are those that 
>insist on trying to force long distance comms 
>through a local area network. 

Ah, but there is *nothing* wrong with "distance"
as long as it is not abusing the network.  PLEASE 
CONSIDER:

1)  WIDE2-2 will generate about 6 typical packets
in this high density area to  generally cover a users
local area.  Most people think that is probbly OK.   

2) The 3 hop path of  DIGI1,DIGI2,DIGI3 that you 
call Long Distance only generates 3 packets.   

Who is placing a greater load on the network?
If all the station needs to do is contact someone
3 hops away in the next town and only takes 3
total packets to do it, then he IS using less bandwidth
than every single one of the W2-2 users.

The stated intent of the NSR routine is to not permit
the user to go 3 hops under any circumstances, and 
will instead blather everypacket to the same area 
(6 digis in this example) whether the sender wants it 
or not.  That is not efficient use of bandwidth either.

In fact,  DIGI1,DIGI2,DIGI3,DIGI4,DIGI5
is a path that would also still use LESS on-air
packets than a single WIDE2-2.  Of course his 
chance of success becomes vanishingly small, 
but still his total load on the network is LESS than
WIDE2-2.  GUARANTEED EVERY TIME.

So again, I think the sender konws best (or should
learn whats best)... and the only thing the network
should do is protect itself from abuse.  And that is
the whole idea of the New-N Paradigm.   Give it 
a chance.  It will be 3 to 5 times better than
before, but still have all the flexibility that anyone
might need.

de Wb4APR, Bob