Hi
I have an 8 channel Lora gateway and wish to have 100’s of nodes that will send a message a few times a day. I understand that the node has a maximum upload messages (and time on air) it can send a day. As each of the nodes we have require to get a very short reply after it sends its message. How does this work with the gateway if there is a limit of number of messages to send a day? If a gateway has 100’s of connections and sends out a reply (from the server application) it will break this rule. How does current systems deal with it?
First of all, gateways do not have connections. LoRaWAN is not WiFi. Devices connect to the LoRaWAN network server, not a gateway.
To answer your question, current systems redesign the requirement for a downlink after each uplink. LoRaWAN is designed for asymmetric communication, from device to the network, not symmetrical.
Why do you need a downlink after each uplink? What happens if the uplink gets lost (or the downlink is lost or can’t be transmitted due to the gateway running out of airtime)?
Keep in mind a LoRaWAN gateway can’t receive when it is transmitting so too many downlinks will inevitably result in additional lost uplinks. Apart from the usual causes of lost uplinks like collisions, RF issues and the like.
As we have no idea what the uplinks are from or why there has to be a corresponding downlink, there is no way to advise if there is an implementation that would work.
As noted above, given the shared use of the ISM band, you may have even picked the wrong communications system all together because:
They don’t because there is no need for the designed use cases for LoRaWAN. Or indeed any user that is using license free radio spectrum as your gateway, LoRaWAN or any other system, will likely breach the legal transmission duty cycle.
However, it is possible to run a very useful scenario with 100s of devices depending on environment and needs.
There are many other technologies that might be better in this scenario.
For LoRaWAN
the gateway listens on all channels simultaneously
messages should be short well under 20 bytes to minimise air time
They should be timed reduce temporal interfere.
you need to stay within legal limits (1% duty cycle) for each device. (non TTN)
That puts you at the 800 device mark maximum (we’ll deal with why you cant do this subsequently)
‘Downlinks’ occupy the channel for much longer (about 6 seconds) and are more prone to interference.
As a result ,downlinks are legally limited to about 10 per device per day
TTN ‘Fair Use’ policy sets usage limits to about 0.017% (approx 1/60 of 1%) duty cycle and is a much better rate to use in designing and operating LoRa systems
LoRa is very good but not perfect at rejecting transmissions in the same time slot and disambiguating one. The other is always lost to that gateway.
Use LoRa as it is designed. A 10 ton truck is not a Ferrari.
Don’t compare LoRa to any other RF technology.
In a densely populated region , other gateways on TTN might help you achieve good throughput. They might also interfere so badly that your system is unusable.
In a rural setting, there may be no other ISM traffic in the LoRa band you are using, your devices might be able to use a substantial part of the total ISM / LoRa bandwidth without interference
This would set you up for 10-100 devices per gateway as a better ratio to use. Device and gateway location needs to be carefully analysed to provide the required RF cover. Shielding by e.g. buildings and mountains can be used advantageously.
A 20 byte payload on SF7 DR5 takes 72mS (uplink only) so 1000 such devices, transmitting once per hour would yield 71 seconds of bandwidth being used in the hour. i.e. about 2% of the bandwidth on one channel. Thats likely to be OK.
If you try with downlinks, 1000 messages take 6000 seconds or about two complete channels of bandwidth. You would need 2400 devices to be legal on a private network (10 downlinks per day). You will receive complaints, closed accounts and possibly a fine.
Assuming no other ISM band traffic, you have a good chance that most messages might get through.
I have tested at half this rate (10000 messages per day) in a rural environment and still lost a few dozen packets every 24 hours.
These tests led me to the conclusion that the TTN ‘Fair Use’ policy is a very fair reflection of the maxima one should design for when using LoRa. 10% of ‘Fair Use’ being an even better target (one message every 30 minutes) where multiple devices are anticipated.
I would not try this anywhere near a major centre since there is usually a lot of other ISM traffic.