If an AREDN mesh has multiple nodes setup to allow access to the Internet, how does the mesh routing algorithm determine which path (node) to use to get to the Internet?
On a related thought, is there a mechanism to specify which Internet route should be the highest priority, and only fall back to another one if the primary one is down?
I'm thinking of a use case where we might have an Internet path setup with appropriate filters etc that we want to be used by default for the mesh, and someone (intentionally or accidentally) enables Internet access through another node.
Really just trying to learn how it all works. :-)
Thanks,
-Dj
You are here
Routing questions with multiple Internet connections
Wed, 01/26/2022 - 11:51
#1
Routing questions with multiple Internet connections
Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer
Dj:
If an AREDN mesh has multiple nodes setup to allow access to the Internet, how does the mesh routing algorithm determine which path (node) to use to get to the Internet?
Lowest ETX.
On a related thought, is there a mechanism to specify which Internet route should be the highest priority, and only fall back to another one if the primary one is down?
Not in AREDN firmware.
I'm thinking of a use case where we might have an Internet path setup with appropriate filters etc that we want to be used by default for the mesh, and someone (intentionally or accidentally) enables Internet access through another node.
Like when someone on the network enables
'Allow others to use my WAN'
and someone connects a Windows computer and that computer decides to do an update?
Please, more detail on your use case where an isolated/island node/node-switch would need internet access.
I hope this helps,
Chuck
Hi Chuck, thank you for the reply. There is no detail, I am simply curious how it works. I am wondering how multiple paths to the Internet would be handled, and if there are any management tools that can influence how that traffic is routed.
By "ETX", do you mean Expected Transmission Count? In other words, least number of network hops?
Thanks,
-Dj
Dj:
By "ETX", do you mean Expected Transmission Count? In other words, least number of network hops?
Yes, lowest ETX.
But maybe not least number of network hops if RF hops are < 100% LQ/NLQ.
RF hops count 1 or more, DtD and tunnels count 0.1.
I hope this helps,
Chuck
-Dj
the expected transmission count definition holds between RF neighbors. But is something different across node hops. Consider a perfect RF link has an ETX of 1, meaning we expect to transmit a packet one time to successfully transmit the data. But if we have a 5 hop of 5 perfect RF links, we still only expect to transmit the data one time to successfully get the data though the network hops, but an ETX of 5 is shown, as if we expected to transmit the data 5 times to be successful.
Also, consider 1 hop of ETX of 5 is not usable, but 5 hops of ETX of 5 is perfectly usable. A combination of hop count and ETX needs to be taken into account to understand what to expect.
Joe AE6XE
Dj,
"have an Internet path setup with appropriate filters etc that we want to be used by default for the mesh, and someone (intentionally or accidentally) enables Internet access through another node. "
One could configure a proxy server, sitting on the LAN-side of an AREDN node and using a separate Internet connection, which you would configure to only allow those who you authorize to connect to it, to route their internet traffic. The users' browsers would then also need to be setup to use this proxy. This bypasses the standard AREDN method for granting Internet access... which in this case should be turned off.
Note that this is intended for the advanced user to configure. To my knowledge there is no best practice documentation on setting up such a device yet.
Andre, K6AH