[robshep]

Instead of a trivial use-case of existing technology it would be far more beneficial to model (1) RF subscriber characteristics and how an unmanaged network of point-to-point devices can support crisis communications when a potentially (very) large number of devices can flip over to long-range point to point. How is congestion managed and mitigated? how is available spectrum, channel space and bandwidth most effectively used? And (2) The paper mentions an upstream message board and the underlying networking features provided by DTN7 describe a range of routing protocols. If nodes are taking part in a self-organising mesh network that can route in and out to central services then the management of layer-1 & layer-2 is crucial to maintain a working network. It doesn't model any of the characteristics of the medium and any strategies for how the underlying physical characteristics can be leveraged to support a mesh.

LoRa is excellent at long range, but with Long Range transmission comes a greater opportunity to interfere with other transmissions. Sure you can reach a long way so a low number of rural casualties can be serviced effectively, but what happens when you have a dense urban scenario and there are 10's/100's/1000's of nodes all hitting refresh on twitter and they are all interfering with each other and the few low-bandwidth gateway nodes are attempting to carry the traffic?

[viraptor]

> 10's/100's/1000's of nodes all hitting refresh on twitter

That's not going to happen with Lora. This it completely different scale of messaging. Your message is limited in bytes and takes multiple seconds to transmit. You can assume the messages will be significantly delayed if 100 people around are actively using it.

But it's designed for a specific purpose and there's no active refreshing/polling that's going to happen.

[robshep]

Exactly. How many people could this support? My guess it would be a small proportion of an urban population.

> no active refreshing/polling

Maybe not twitter specifically, but the paper mentions a centralised message board like twitter, so we can assume some sort of interactive service would be suggested.

My main point is this that for such a specific bearer (bandwidth, channel plan, duty cycle, range etc) it is important to model the capabilities. I’m not sure how well this would support a crisis (notably free of central management) when there is the potential for large numbers of nodes.