[aprssig] Excessive beacons/connects by "smart" IS clients
Lynn W. Deffenbaugh (Mr)
ldeffenb at homeside.to
Sun Apr 15 15:33:06 CDT 2012
There is absolutely no reason, IMHO, for connecting and disconnecting
for each beacon. APRSISCE/32 establishes a connection to an APRS-IS
server and maintains that connection until an error is encountered, it
goes quiet too long (missing heartbeats indicating a non-error-detected
loss of connection), or the client is closed.
If there's an APRS-IS client out there that is connecting, beaconing,
and disconnecting, then I would consider that to be a busted client
which needs to be fixed, regardless of the beacon frequency.
Lynn (D) - KJ4ERJ - Author of APRSISCE for Windows Mobile and Win32
On 4/15/2012 9:11 AM, Gerhard.F5VAG wrote:
> The various APRS-IS servers are exposed to an increasing number of
> mobile TCP/IP clients which tend to send at beacon rates exceeding
> reasonable limits. This issue will rapidly increase with the increasing
> number of "smart" portable devices.
> In these cases aprs.fi sends the kind warning
> "This station is transmitting packets at a high rate, which can cause
> congestion in the APRS network.".
> However, seen from the server to which the client is connected, the observation
> is much worse. In certain cases the server has to handle connect/disconnect
> cycles below ten seconds over considerable time periods:
> Many of the excessive beacons are fortunately dumped by the entry point
> server and not forwarded to the main stream, e.g. dupes.
> In my view it is not reasonable to sent beacons with an interval of less than
> five minutes with TPC/IP connections, but to use UDP. Many T2 servers provide
> this possibility on port 8080. It drastically reduced the overhead for
> the servers
> and the clients by avoiding the connect/disconnect overhead.
> Even for clients using UDP the beacon interval should be limited to 30
> seconds. I know that this limits the resolution to 0.8 km @ 100km/h
> (0.5 miles @ 60MPH) and think this is a reasonable limitation.
> In any case, more "smart" beaconing should be used..
> Any comments?
> 73 de Gerhard, F5VAG
> http://f5vag.eu, france.aprs2.net
> P.S.: The callsign in the above links is certainly not real.
> aprssig mailing list
> aprssig at tapr.org
More information about the aprssig