Number of nodes in LoRaWAN

ChirpStack Network Server
1

I am new to LoRaWAN, so my question may seem little absurd.

My question is, How many nodes can remain practically attached to a loraserver?

2

As many as your hardware supports connecting at the same time. Since most LoRa nodes will spend the vast majority of their time asleep and only check in periodically, scaling is based on concurrent check-ins and not maintained connections.

3

This is a serious misunderstanding of how the pieces of a LoRaWAN system fit together.

LoRa hardware has no concept of a unique node or a “connection” - only the server even has any idea of that. Most packet forwarders don’t even know if the traffic they are passing is actually LoRaWAN - they just know that that they have some bytes received with a valid hardware checksum.

Since most LoRa nodes will spend the vast majority of their time asleep and only check in periodically, scaling is based on concurrent check-ins and not maintained connections.

That’s not entirely in conflict with the truth.

But it would be much simpler and more accurate to say that it depends on how active the nodes are.

4

Feel free to re-read my post at your leisure. Obviously “hardware” was a reference to the server side accepting payloads. It makes little sense as read the way you did, as most (if any) LoRa devices will not be joining more than one network server.

5

That makes no more sense than the original erroneous claim.

Any limit in the server would be one of software allocations. In practice, even if there is a design limit in a server, on-air contention from the uncoordinated nature of LoRaWAN nodes is almost certain to be the more meaningful limitation, though one that we both seem to agree scales with their activity.

This has nothing to do with multiple networks, though having multiple networks in the same geographic / frequency space will practically reduce the data capacity of each, as is true in any situation of bandwidth sharing.

6

I think there are two bottlenecks:

  • LoRa gateways (e.g. the packets that can be received simultaneously by a gateway)
  • Server hardware (which you can scale up / down as you need)

This might be an interesting whitepaper:

1 Like