[aprssig] Old Digi_Ned Refurb

Jason Winningham jdw at eng.uah.edu
Thu Sep 13 18:23:22 EDT 2007


On Sep 13, 2007, at 3:27 PM, Robert Bruninga wrote:

> But it does the job very well.

at layer 0 and 1, yes. As a network infrastructure device, not so  
much so.

> Well, the second biggest load on the APRS network prior to the
> New-N paradigm was the flooding of the network with too many
> digipeater position reports over too large an area too often.
> Switching to proportional pathing to reduce the rate at greater
> hops was a real reduction in QRM.

Doesn't that make the local repeater initiative a big problem, then,  
since it uses the digi's slots that could otherwise be used for  
proportional pathing?

> Such a setting would worry me significantly.

Why?  IP does this, and we all know how unreliable the IP based  
internet is...

> To be reliable on
> demand, APRS must have consistent performance expectations for
> users and not be at the whim of each digi owner to be changing
> the intent of the sender,

The capabilities of the network are not dependent on the intent of  
the sender.

> The New-N paradigm has gone a long way towards bringing
> consistency to the network without resorting to draconian and
> arbitrary modification of user packets by individual digi
> owners.

No, it resorts to the draconian and arbitrary method of replacing an  
Y hop path element with a one hop path element (if your TNC has  
enough alias slots left open).  What's the difference?  I the user  
_need_ my packet to go 14 wide hops, so who are you to only allow my  
WIDE7-7,WIDE7-7 to only go two hops?

If proportional pathing is so great, why not implement it at the digi  
so it doesn't matter what the "dumb" tracker setting are?  The digi  
could say "hmm, this guy got three hops last time, but that was 62  
seconds ago so he gets one hop this time."


Argh, I should keep my mouth shut (er, keyboard idle) and not let my  
self get sucked into commenting, but networking is a noticeable part  
of my day job and the serious lack of capability in the APRS  
infrastructure bothers me.

-Jason
kg4wsv






More information about the aprssig mailing list