Homework 4: Network Layer

CNT 4007C Computer Networks Fundamentals

Homework 4: Network Layer

(submission instructions similar to hwk3, on canvas then hard copy in class and ofc hrs)
Q1. What are the 2 most important network-layer functions in a datagram network, and
what is the difference between them?
Q2. How many columns does a bare-bones forwarding table in a MPLS-capable router
have? What is the meaning of the values in each of these columns?
How many columns does a bare-bones forwarding table in a datagram network have?
What is the meaning of the values in each of these columns?
Q3. An application generates chunks of 40 bytes of data every 20msec, and each chunk
gets encapsulated in a TCP segment and then an IP datagram.
What percentage of each datagram will be overhead, and what percentage will be
application data?
Q4. What is the triangle routing problem in mobile IP? Suggest an improvement to it.
Q5. Detail two examples in which tunneling helps in mobile IP and IPv6 deployment.
Q6. How does IPv6 help in supporting mobility, compared to IPv4?
Q7. Why is MPLS sometimes called layer 2.5?
Q8. Can label numbers be reused in MPLS? Why?
Q9. Mention two examples in which IP addresses can be reused, explaining the reasons
for their reuse.
Q10. Compare and contrast link-state and distance-vector routing algorithms.
Q11. Compare and contrast the advertisements used by RIP and OSPF.
Q12. Fill in the blank: “RIP advertisements announce the number of hops to various
destinations. BGP updates, on the other hand, announce the _____________ to the
various destinations”.
Q13. Complete the following sentence in three different ways, each time using one of the
following words: ‘vector’, ‘domain’, ‘gateway’
“RIP is a ______ ______ protocol, while BGP is a _______ ______ protocol.”
Q14. Suppose you purchase a wireless router and connect it to your cable modem. Also
suppose that your ISP dynamically assigns your connected device (that is, your wireless
router) one IP address. Also suppose that you have five PCs at home that use 802.11 to
wirelessly connect to your wireless router. How are IP addresses assigned to the five
PCs? Does the wireless router use NAT? Why or why not?
Q15. Consider the network setup in the figure below. Suppose that the ISP instead
assigns the router the address 126.13.89.67 and that the network address of the home
network is 192.168/16.
a. Assign addresses to all interfaces in the home network.
b. Suppose each host has two ongoing TCP connections, all to port 80 at host
128.119.40.86. Provide the six corresponding entries in the NAT translation table.
Q16. What is the ‘rendezvous problem’ in multicast? How can it be solved? (mention
three main approaches/algorithms to the solution along with the protocols that use them)
Q17. How does SDN extend the notion of matching to define firewalls and NAT boxes,
while traditional longest-matching is limited?
Q18. Do you see a layer violation in BGP peering? Explain.
Q19. Provide example of route oscillation in link state routing (e.g., Dijkstra’s algorithm).
Q20. Explain the ‘count-to-infinity’ problem and why it happens in distance-vector
routing algorithms (like Bellman-Ford in OSPF), then suggest an improvement. Mention
a limitation of your suggestion.
Q21. What is the advantage of using multicast over unicast? Give example and provide a
quantitative analysis (i.e., numbers) to solidify your argument (you can use a drawing of
the network topology).
Q22. [Extra points] Consider the network topology in the figure below with nodes x, y, z.
Assume that distance-vector routing is used with Bellman-Ford algorithm. The link cost
between x and y changes from a to a’ , where a’ a, and a’ c. Focus on nodes y and z,
their routing table updates, and messages exchanged between them.
I. Derive an equation for the number of iterations needed for the routing to
converge as a function of the link costs a, a’, b and c.
II. Suggest an improvement that would speed up the convergence and prevent
looping in the above case. Explain your solution and how it works.
Q23. [Extra points] Mention the two main reasons why broadcast-and-prune multicast
does not scale.
Q24. [Extra points] Mention two reasons why an optimal RP placement is not used (or
needed) in PIM-SM.