[aprssig] time for APRS second generation network?

[Jason Winningham]

> By the way, Iam not at all opposed to  you sugggestions
> for a smarter infrastructure.  Dont get me wrong.
> I am all for a 9600 baud RF backbone system say
> with a state (or ARRL section) that makes a great
> LAN boundary)...

I haven't studied the 9600 baud backbone system, but at a glance I 
can't see how that will help local nodes reliably send and receive data 
amongst themselves.  as far as I can tell the 9600 baud backbone can 
improve send to outside, can improve receive from outside, or you can 
make your network bigger.  If you improve send OR receive but not both, 
you have an asymmetric data flow and someone is not getting the whole 
picture.  If you make your network bigger, well, it's already too big.

As far as I can tell, the only way for 9600 baud to improve "A tactical 
real-time local ad-hoc network for
digital communications among everyone..." is for _all_ nodes to switch 
to 9600 baud.

> Im just saying that we MUST fix this level 2 saturation
> problem first.

I'm not saying we shouldn't do anything now, or that there's anything 
wrong with the work being done now.  I'm proposing the next step, which 
is to get a layer 3 with some smarts because we are currently 
demonstrating that layer 2 cannot do enough, especially the layer 2 we 
have now.

>   and the smarter infrastructure cant do a
> anything about that.

Sorry, but that statement is wrong.  I began my first message with an 
example of a real network that had everything on a single layer 2 
ethernet, but no layer 3.  Doing nothing more than adding layer 3 
(replacing some layer 2 devices with layer 3 routers) improved that 
network orders of magnitude.  The same basic networking principle _can_ 
improve the APRS network.

Because of the nature of our physical layer we cannot see same same 
"orders of magnitude" improvement, but we CAN:

- simplify user configuration
- adjust the local segments of the network to handle traffic 
appropriate to the local network load
- reduce duplicate repeats
- reduce path lengths, both number of hops and physical packet size

>  And I just want to un-confuse
> users.

Yes, and I think it should be whether they want it or not; that is, let 
the network infrastructure make decisions to benefit the whole network 
rather than being at the mercy of any ignorant or malicious user due to 
this archaic source routing scheme we use now.

> The bottom line is we MUST cut path length and
> high-digi coverage

Right, but these are two completely separate issues.  Nothing we do at 
layer 2 or above can have a useful impact with a very badly configured 
layer 1, like a digi antenna that can see the entire northern 
hemisphere.

In a nutshell:  yes, there is a problem, yes New N-n helps if everyone 
plays, but what are you going to do next time, when New N-n has reached 
its limit?

-Jason
kg4wsv