So I have had a thought about this. A long and hard, critical thought
This network is no different to the former WUGs.
It is not commercial and therefore is really best effort
This network will provide you a best-effort means to get your widget's signal from one point to another, so your widget, we could call an "entry node", sends things into the mesh, but you would need an "exit node" somewhere to connect whatever your widget is sending, to either an application or a server.
So this will likely not be the kind of thing you would use for these use cases:
* Alarm monitoring (note NOT armed response as that is subject to rules and regulations)
* Sensors with kind-of-sorta realtime performance, even a low sampling rate of one sample per 10 minutes with a lot of sensors are going to flood this network as we learnt by our own implementation of a mesh at work on our products.