Historically, since we have apps parked in that spot, there has been
an insistence that all of the topology complexity get resolved in the
lower layers.
uh, no. the design is for a clean separation of function between the
layers that know about network topology and the hosts, which aren't
supposed to need to know about such things. that separation of function
is still highly desirable.
what has changed is that we're now expecting a layer 3 that was designed
for relatively stable networks of wired hosts to suddenly accomodate
mobile and nomadic hosts and networks, without significant change.
you can't solve the right problems without first recognizing what they
are.