On Feb 19, 2015, at 1:29 PM, Ralph Droms <[email protected]> wrote:
> If I can extrapolate and oversimplify a bit, now we've gotten to a 
> fundamental problem: how does a random collection of devices, links and ports 
> sort itself by DWIM into a coherent home network?  How does a device with 16 
> ports decide to group ports (0-5, 7, 10), (6, 8, 12-15) and (9, 11) into 
> separate subnets?

What Mikael is saying is that on homenet devices (not switches) each port is 
treated as a separate network, on the assumption that all of the physical ports 
on the homenet device are likely to connect to routers, and that any switched 
topology would be done using switches connected to homenet routers.

I am not convinced by this; one of the issues I see here is that until DNSSD 
comes up with a solution, we are breaking service discovery on the home network 
by doing this.   Granted, we already do that from Wifi to wired, but this seems 
gratuitous.

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to