Surely. The problem is that sometime you’ll be unable to control them too.
Yes it is unrealible by design but in the last 8 month, since my gateway and some nodes are running, it happens much ofter that the TTN Backend has a problem (my monitoring alerts me) than a packet from a node is not received by the gateway. I can live with this sometimes instable backend, my use cases working anyway, but i guess in a industrial environment it is necessary that you need a stable running backend on that you can trust it is working.
Hello my friend, I’m from Gelsenkirchen from the Ruhr area 
Yes, I have that feeling too. In these cases, I would connect the gateways with TTN Broker, but at the same time to its own? So you are not dependent on TTN? Is that possible?
We will use LoRaWAN for Smart City, but we also want to support the community.
It would be more for sensor values. So trash can, air quality, parking …
Different SF are “almost orthogonal”, not “absolutely orthogonal”.
Readers interested in this should refer to:
“Impact of LoRa Imperfect Orthogonality: Analysis of Link-Level Performance” by Croce et. al.
“On the evaluation of LoRaWAN virtual channels orthogonality for dense distributed systems” by Ferrari et. al.
and
“Mathematical Model of LoRaWAN Channel Access with Capture Effect” by Bankov et. al.
3 Likes
not much to do with the topic title but ok
here’s one too All About LoRa and LoRaWAN: LoRa: Orthogonality
1 Like
From another topic:
At around 14"20’, transcript and emphasis mine:
The long range is actually also a disadvantage of LoRa. Why? Because when you have a receiver which receives in a 2 km area, if you put a lot ot sensors there, let’s see, if you put more than 2,000 sensors in this range of the transmitter, they almost cannot connect to the gateway because everyone attempts to talk and jumps with somebody else who also wants to talk. So this is one of the problems with LoRa.
In Plovdiv we don’t have this problem. But in Sofia there are a lot of people who try to implement LoRa solutions. And, you know, they work on the license free frequency. But “license free”, this doesn’t mean that this frequency is yours. If you want to follow the standard, you should not talk more than 1% of the time. Nobody complies with this. They just talk until the gateway confirms that it received the message. And they make it impossible for other people to follow the standard.
So I had a customer last week and they said “show me where this LoRa works, because we tried in Sofia and it doesn’t work”. I said, yeah, here in Plovdiv it still works, because there are not so many users who do not follow the standard and just emit content, maximum possible distance, the longest messages they can do, and don’t care about the others.
Of course, the following map only shows TTN gateways, but: I wonder if the above can be mitigated by installing more gateways, to decrease the cell size along with ADR:
When one signal is a lot stronger than the other, than it can survive a collision like mentioned above.
1 Like
I watched/listen to that presentation also and thought much the same - per my comment in the linked thread many focus on just doing nodes and forget to add to the pool of GW’s. Also the map is typical of TTN deployments in that # of GW’s could be increased 40% if the offline GW’s were left on/connected
With brilliant sense of timing typing this just interrupted by UPS Guy delivering YALG* for me…#26/27 here we come 
*Yet Another Low-Cost GW
I would think that more gateways and working ADR would go a long way towards solving this problem. Of course spectrum expansion would also help. Right now LoRa has a very small space to work within.
Or maybe not.
For years a solution to traffic congestion has been to build more and better roads.
2 Likes
Interesting case from Germany: A non-TTN node was sending every 6 seconds with SF12 and effectively blocking the nearest gateway. OTAA was nearly impossible.
The affected person contacted the Bundesnetzagentur (https://en.wikipedia.org/wiki/Federal_Network_Agency) who is responsible for radio frequency violations and they quickly responded with a local investigation and found the rogue node after approx. 1.5h of search. They are now investigating if this is illegal in Germany (the local investigators did not hear of LoRa up to date point) and the user tried to reduce the sending frequency (he probably did not know what he was doing).
3 Likes
Now imagine that some nodes continiously transmit just LoRa preamble. You’ll see no uplinks from other networks in this case, and you gateway will be effectively blocked. Are you going to ask the GeheimStaatSpaceMarines to come and rescue you every time when you encounter “no uplinks in TTN console” trouble?
A non-TTN node was sending every 6 seconds with SF12 and effectively blocking the nearest gateway. OTAA was nearly impossible.
It shouldn’t - unless your scheme only supports one frequency, or unless that node is using the same frequency selection plan you are, and you spend periods of timing being synchronized in your execution of such.
Now imagine that some nodes continiously transmit just LoRa preamble.
That would probably be illegal in most places.
First, you generally aren’t allowed to transmit narrow bandwidth signals continuously. Maybe you mean “frequently”
You’ll see no uplinks from other networks in this case, and you gateway will be effectively blocked.
Not really. One pre-amble sender would only hit one frequency and spreading factor. A gateway not close enough to be swamped outright or picking up the skirts of the signal should be able to receive on the other channels and even the other spreading factors of that channel.
Are you going to ask the GeheimStaatSpaceMarines to come and rescue you
In most places it is illegal to intentionally interfere. A system just playing “made you look!” with LoRa demodulator slices by sending preambles but never following with any data would likely be hard to defend as anything other than an intentional interferer.
1 Like
One SF12 uplink every 6 seconds is illegal in EU too; at all, we’re talking about misbehaviour in this topics, aren’t we?
It will hit one physical demodulator. There are eight ones in SX1301. Just do try
More interesting is that in EU with 10 devices you can more or less legally block RX2 for all the devices around; bye-bye class C
The demodulators aren’t as fixed in purpose as you seem to think. There are actually at least two stages with a rather fancy dynamic handoff assignment between. You can only receive 8 signals on the flexible paths, but with the right spreading factor combinations those 8 don’t have to all be on unique frequencies.
More interesting is that in EU with 10 devices you can more or less legally block RX2 for all the devices around; bye-bye class C
There are costs to LoRaWAN’s decision to forgo short-term, repairable state such as could permit synchronizing a hopping schedule for class C after a node does uplink. Ironically, where LoRaWAN does use shared state (to prevent join nonce or fcnt re-use) it tends to be all but impossible to restore once lost. It’s repairable short term state that is key to how many other sorts of modern multi-frequency networks work.
You’re pretty right. So what? Eight preamble sender will hit all the available demodulators, will not them?
8 boxes sending just preambles are obviously maliciously interfering. While this thread is about “guidelines” LoRaWAN is mostly used in places where there are “laws” prohibiting such behavior.
Please take a look at the very first post.
Isn’t this prohibited by law in most places?
where no plaintiff there is no judge.
And even you have to find the node.
If some hackers will disturb the network they do it.
Those same hackers can get a cheap box to interfere with mobile phone signals and interrupt coverage at a location. That is the nature of wireless transmissions.
3 Likes