> 1) I'd call that a "leaf router" (or maybe "branch router") model, as
> it assumes symmetric routing, which isn't a reasonable assumption in
> the presence of redundant paths.

However, it *is* valid for many routers in many networks of reasonable
size.

> 2) "leaf-router" vs. "strong host" vs.  .. status should be on a
> per-interface basis.

As I'm sure Håvard knows, this is precisely the functionality you get
with the per interface "ip verify unicast reverse-path" functionality
in some versions of Cisco IOS.

Having such functionality available is definitely a good thing.

Steinar Haug, Nethelp consulting, sthaug@nethelp.no