2) Starting from bootup, for a 15 node ring
network, how long will it take for all the nodes to
have the same routing state if updates are done
every 10 seconds and a distance vector protocol is
used. How about for a link state routing
protocol?
3) I mentioned in class that one of the
improvements in RIP 2 is password authentication and
we talked about injecting false topologies.
Can you think of another way to protect from someone
injecting false topologies if for some reason
password protection was not feasible?
4) Why do link state routing protocols use flooding
when each router typically has a complete network
map and could send LSA updates directly to each
router in the network?
5) Some network architectures include separate data
and control connection, so that the traffic can be
kept separate. Why would this be an advantage? Of
the 4 main routing protocols we discussed in class,
which would benefit the most from a network with
separate data and control connections? Are there
security implications of such a network?
6) If delays are recorded as 16-bit numbers
for a 100 router network and delay numbers are
exchanged twice a second in order to optimize
routing, how much bandwidth per line is used up by
the delay component of the distributed routing
algorithm. Assume each router has 3 links to
other routers.
7) In class, we talked about how to use
Dijkstra to find the shortest path through a
network. Describe how to use Dijkstra to find
two path through a network such that if any link
failed on one path, the other path would work.
You can assume that routers don't fail, only links
between routers fail.
.