I have the same question too.
I’m using the Semtech Packet Forwarder to connect to TTN Stack v3.
And I’ve setup the “Location source” with the “Update from status messages” option.
But the location of the gateway can be updated automatically even the “gateway status” was forwarded to TTN.
Like the posts above, it has “Update from status messages” enabled. But note the help text:
When checked, the location of this gateway will be updated from status messages. This only works for gateways connecting with authentication; gateways connected over UDP are not supported.
TTN Mapper is showing it nicely though:
Regardless, I now just disabled that checkbox and copied the coordinates and altitude from the message into Console, as it is in a fixed location anyhow.
Not allowing unauthenticated UDP packets to update the gateway location might be considered a security feature as there is no way to validate the information is provided by a gateway itself and not by an imposter. (Which happend a few years back)
True, but even the user can simply enter incorrect position data!
So how likely is it to transmit position data incorrectly, and what impact does it have on the functioning of the network?
It’s not about the gateway sending wrong information itself (or the owner making a mistake), it’s about someone using a script to set wrong location for other people’s gateway(s). Which has happened.
Oh, what are the chances! It seems somehow TTN’s Slack is running a free trial right now, making all old messages temporarily available again. (Also PMs, even if you start a “new” PM thread with someone not in the list of recent PMs.)
The massive fake locations happened April 5th, 2017:
@htdvisser etc, it seems like a lot of the gateways are reported to be at location “gps”:{“latitude”:50.008724,“longitude”:36.215805} by the noc.
Seems like it started happening at 2017-04-05 12:00:31 UTC
Jumping around between the correct and the false location.
Perhaps someone in the Ukraine found a way to send status messages with a wildcard gateway EUI?
how many is “a lot of the gateways”?
Because we’ve seen some cases where different gateways use the same EUI, and depending on the last one to send a status message, the coordinates change
And does this only happen in the data the NOC returns or also in the data ttnctl returns?
We’d love to drop the Semtech protocol. It’s unreliable, insecure and causes many many problems on our servers, but unfortunately it’s installed on virtually every LoRa gateway, so we can’t drop support any time soon
The behaviour that I have seen with the “Update from status message” setting was slightly different. Gateways running UDP Semtech forwarder:
Setting off + location set on console: manual configured location is available on mapping APIs.
Enable setting: the location of the gateway is set to 0,0 automatically and the mapping APIs reports the gateway at 0,0
Disabling the setting requires to set the gateway location again.
This does feel like a bug, but maybe the “clearing” of the location when this setting is enabled is intended.
The current behaviour is a little unintuitive, but I don’t mind having to set the gateway locations by hand on the console. Only using manual set locations can be more reliable as one eliminates some variables like GPS drift.