In message <CADiurz38Hq4+xxYhx+s7XDOOv5rsaKxGnNj=0w68p3zg4jm...@mail.gmail.com> Victor Kuarsingh writes: > > > > And if I'm at home with WiFi at home and my laptop plugged into the > > Internet and plug in the USB (or other tether) to update my contacts > > list on the phone, because its easier/faster, we have the situation > > described, even though the intent was not to use the phone as an IP > > gateway. > > I can see your point. So in this case.. I guess your computer/laptop > would belong to two networks (wired USB and Home Network over WiFi). > Not sure if this example shows how the two networks join - unless the > computer bridges/routes traffic.
Good point. The computer would have three ways to pick from. Maybe the phone acting as a WiFi was a better example. > > This gets back to users plugging things in in seemingly arbitrary way > > that made sense to them at the time for some reason. > > Would you say this example adds merits to ND for the smartphone > gateway function, or the opposite? ND just finds the link local addresses. Figuring out which way to go with packets is another story. You need a routing protocol if there is a loss of connectivity more than one hop away. At that point the addresses are still valid, but the path to the default route has changed. There is no fast withdraw in ARP proxy, DHCP proxy, ND proxy. In a routing protocol, the flooding of bad news is fast. Plus arbitrary topologies can be supported. And as Jim pointed out, we need to be careful about the scalability of flooding for WiFi networks that may be come dense. > regards, > > Victor K Regards, Curtis _______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
