On Thu, 14 May 2009 10:04:21 -0700, "John Hays" <[email protected]> said:
> If the controller firmware would pass "ALL TRAFFIC" to the gateway the  
> whole "G" port thing could go away.   A smarter piece of gateway  
> software could determine what needed relayed somewhere else and what  
> is just local repeater traffic.

Let the router be the router, eh?  That's a great idea, John.  Wish that
could happen.

> A) Radios (RF) on D-STAR did not need to register - let the gateways  
> just update current location into a dynamic, distributed data store.  

Agreed.  I always assumed "registration" was to meet regulatory
requirements somewhere, but the more I thought about it, regulations are
written that each Amateur Station is responsible for their own
transmissions.  If anything, the Gateway should have had a "blacklist"
for protection of the operator of the repeater from malicious
interference or other issues -- but as we thought about that... ANYONE
can become "WY0X" or whatever callsign they want at the spin of a dial,
so filtering by callsign is completely "busted" as a concept right from
the start.

> (FIlters at the gateway could manage who could and could not use a  
> given gateway - and callsign pattern matching, e.g. regex, could  
> filter out most bogus callsigns)

I don't like the idea of filtering "bogus" callsigns.  What might be
"bogus" to you, might be my special event's tactical callsigns. 
(There's nothing stopping anyone from registering "SAG1, SAG1, NET,
EVENT", etc.)

> B) Only register "network devices" such as gateways, reflectors,  
> dongles, etc. and have things like dv dongles, autopatches, cross  
> service links (e.g. IRLP), run through servers that present them to  
> the D-STAR network.  Redundant "trust servers" could maintain this  
> registry to remove the single point of failure.

Hmmm.  Thinking about that one.

> This would facilitate such things as mobile D-STAR stations moving  
> from repeater-to-repeater or port to port in real time.
> Enable addressing services (reflectors, dv dongle type devices, nets,  
> etc.) directly in "UR" field.

Real time roaming certainly seems like it should have been "do-able"
right from the start... 

> RPT2 could be used to route to a different repeater/port and eliminate  
> the whole "/repeater" syntax.

Another interesting thought.  

Thanks John.

Nate WY0X
--
  Nate Duehr
  [email protected]

Reply via email to