[Xastir] UDP frames

[Alan Crosswell]

Don't broadcast it, multicast it.  No need for uninterested hosts to have to 
receive your APRS packets.
/a

Wes Johnston wrote:
> Perhaps we've meandered far enough it two days....
> 
> My goal when I brought this thread up was to create some widget (or a new port
> type under xastir) that would connect to the aprs IS stream and pass that data
> out on my local wlan (or lan for that matter) as UDP packets on port 14439.  In
> the case of a wlan, the ssid could be APRS or AXU.  The point to this is that as
> my laptop/pda/whathaveyou wanders in and out of hotspots around an event, I get
> any and every packet that is available from the AP that is broadcast to
> 255.255.255.255:14439.
> 
> If we start talking about TCP packets, I may as well establish a connection to
> some normal APRS software like xastir on TCP port 2023.  Problem is, in a local
> event, there may be long periods of no traffic and I could not use (as a client)
> the auto reconnect feature of most aprs programs because they would not know the
> connection had dropped due to my wandering around, or if there had just been no
> data for an extended period.  Even then, I have to reconnect and during the
> unconnected time, I missed some packets.
> 
> I like the UDP method because there is no connected layer overhead, no ACKs
> (well ok, there are ACKs in the 802.11 mac layer).  In the end, the software on
> my PDA just listens promisculously on port UDP 14439 and scavenges anything it
> can.  This is no different than APRS listening to PID F0 (UI frames).  There is
> no setup of a connection that may fail in 2 seconds as I pass out of range of
> the 802.11 ap... just catch as catch can.
> 
> I understand that 802.11 works a behind the scenes connection when you wander in
> range of an AP... But I don't need to know about that in this context... all I
> know is that I wander within range of a hotspot and like magic data starts
> appearing on port 14439.  My not worrying about the behind the scenes
> connection between 80211 devices and the AP, I closely mimic being connected to
> a LAN cable and hearing the broadcast packets there.
> 
> Also by using UDP frames, we don't have to know which host to connect to...
> which is a pretty major part of APRS... come as you are and hit the ground
> running.  If I run TCP telnet connections to a host, I have to know the name of
> the host, and program that into my client software.  A difficult thing to
> convey.
> 
> I did like the analogy someone said about working packet this way would be like
> meteor scatter.  That's true when driving a car.. but in/on/around the EOC,
> anyone with a laptop in range of an AP could have the full aprs IS stream w/o
> having to do anything other than turn on the network card.
> 
> I need to check Curt's suggestion of something simple like Netcat.
> 
> Wes
> --
> 
> 
> 
> Quoting jeff at aerodata.net:
> 
> 
>>Yes, now you are seeing the problem. I need to get in there, assoicate,
>>get an IP address, and then dump my data to the APRS server stream.
>>
>>In the later case, making a more terse method to enter position reports
>>makes sense (if it already doesn't exist).
>>
>>To associate quicker, you need to step up the 802.11b probe request rate.
>>
>>The IP address, not much we can do about that as that depends on the DHCP
>>server. But we have control over the probe requests as well here.
>>
>>If you run windows, I'm sure you have heard of netstumbler. Its a AP/Adhoc
>>scanning tool. Similar things exist for linux. What you may not have
>>heard, is netstumbler has its own API...
>>
>>http://www.stumbler.net/scripting.html
>>http://www.netstumbler.org/index.php?
>>
>>Which should allow you to use the GPS and AP data to more quickly connect
>>and dump data to the APRS system.
>>
>>
>>
>>
>>>>What if you are running adhoc?
>>>>
>>>
>>>That's was the whole point of my earlier questions.
>>>
>>>I might be romaing around at 60MPH and see any given access point for only
>>>seconds - if at all.
>>>
>>>I am interested in how connectivity will work with an access point at a
>>>high
>>>location and a beam pointing down at me. Over all, I suspect the whole
>>>deal
>>>might look more like meteor scatter than packet radio :-)
>>>
>>>Andreas, N6NU
>>>
>>>_______________________________________________
>>>Xastir mailing list
>>>Xastir at xastir.org
>>>https://lists.xastir.org/mailman/listinfo/xastir
>>>
>>
>>_______________________________________________
>>Xastir mailing list
>>Xastir at xastir.org
>>https://lists.xastir.org/mailman/listinfo/xastir
>>
> 
> 
> _______________________________________________
> Xastir mailing list
> Xastir at xastir.org
> https://lists.xastir.org/mailman/listinfo/xastir