How would Homey respond to being connected to a network that has redundant internet ?
IP and Gateway and such would remain the same. I would just setup a second WAN port on the router. Router does NAT so the Homey itself would not see anything different.
Main question is, would I still be able to reach Homey from outside the house or would the Atom server loose connection to the Homey and should I wait for Homey to reconnect to the server before I could reach my Homey again ?
2nd WAN would be a Starlink in ‘Standby mode’ and WAN config would be ‘failover’ due to the Standby mode haveing a small bandwidth…
You would think it would work automatically, but given that Homey’s cloud management is notoriously bad (it’s the reason why many people get “Homey offline” errors even when it’s not), you should properly test this.
For instance, I don’t know if Homey will be able to detect that its previous connection to its cloud servers has changed. I also don’t know how Athom’s cloud servers will handle the change of originating IP-address.
That’s the point, I don’t think this will work automatically, hence the question…
I do suppose having a backup WAN would help Homey in maintaining connection to Cloud devices, since it doesn’t know through wich connection it has acces to the cloud.
But beeing able to connect to Homye would be a great plus…
That’s what I’m trying to say: even if it works for someone else is not a guarantee that it will work for you, because of the issues that Homey’s cloud implementation has.
If you know how to fence it of, you could enable port forwarding. Then all you need is to remember both public ip addresses and switch manually in your app.
Before we got our fiber connection we had a dsl (2mbit) connection with a ‘KPN Buitengebied’ 4G connection for a couple of months. The dsl connection was crappy and the 4G regularly went offline for a bit so our internet connection switched very often between the two. (Totally different public IP’s)
I haven’t actively monitored Homeys connection but I also never experienced it wasn’t available, so I’d say the session was persistent in the way any disruption was to short to notice.
