We are developing a updated version of our product that is currently using the Photon. A customer could have multiple devices in their facility, and right now they have to setup WiFi on each product in a facility that might not have WiFi reaching every device perfectly (or ethernet run to every device). The idea of a mesh network is ideal for us.
I was wondering how the Argon’s handle multiple gateway capable devices. Ideally we want as little effort for our customers to install and begin operation of their devices, so hard coding what mesh network the devices will connect to is idea (I can’t see any way right now to set up those credentials in software).
Ideally, if multiple Argons in the same area, one would act as the gateway and the rest communicate over the mesh, but I am not sure if this is how it works or if it could be setup to work like that. We would like to be able to ship every device to work as a gateway in case they only have one, but act on the mesh if its available with little input from the end user.
Just to get a feeling for your request, how do you weight the cost of a packet going out via a “secondary” gateway vs. routing cost (i.e. additional latency)?
If you have multiple WiFi gateways, any mesh packet would normally take the “quickest” route - commonly to the nearest active gateway, unless congestion causes a longer route - reducing node hops and latency cost.
Having multiple WiFi or Ethernet gateways would commonly valued at zero cost, since the mere registration of the gateway (independent of active uplink state) would already make the difference between a micro network (price tier 1) and a high availability network (priece tier 2).
Our priority is reducing monthly cost, but allowing devices to receive remote firmware updates. We are not doing much data reporting, so latency doesn’t cost us much.
Our big issue is every device needs to be capable of being the gateway, but because we could have multiple products and multiple types of products in the same building, auto negotiating who is the gateway would be super powerful.
Again, in what way would you anticipate any cost savings by not having a gateway online.
Currently I don’t see any way to change a gateway (connected or not) into a mere router or edge node on the fly.
But if this would be a widely needed feature filing a feature request on GitHub would be an option.
I am not sure what the price is for a Large Site Network (which would be any mesh with more than 3 gateways online - which would cover our application) would probably be more expensive than the High Availability network (which is more expensive than the single gateway pricing). If we need every device to be gateway capable, sticking with a Photon is much cheaper: 0.39 per device per month versus 2.99 per micro network (one network gateway)