CS 6543 Computer Networks

Spring 2005 – Midterm

March 10, 2004

Name:……………………… Score: ……./20

This exam has 6 questions in 8 pages. Good luck.

1. (2 points) In this question, you will explain/discuss some concepts briefly.

a. (0.5pt) What are the key differences between link-state and distance-vector routing?

b. (0.5pt) Briefly illustrate/describe how a system admin can use ICMP protocol with
redirect messages to dynamically update routing tables of hosts.

1
Name:………………………..

c. (0.5pt) What is the purpose of DNS? Briefly describe its operation.

d. (0.5pt) Briefly describe how to acquire network map in a link-state routing protocol?
(give an example)

2
 Name:……………….
2. (4 points) Consider the below network. Suppose Node A generates a 1-KB data packet every x
msec and tries to send it to node D. Each link has the same propagation delay of 1msec. The
capacity (bandwidth) of A-B is 4 Mbps (mega bits per second), that of B-C is 2Mbps, and that
of C-D is 1Mbps. Assume that packet processing time is zero at every node. Assume K(kilo)
=103 , and M (mega) = 106.

4Mbps 2Mbps 1Mbps

A B C D

a) (2pt) Assume that x is 5msec(i.e., node A generates the first packet at t=0msec, the second
one at 5msec, the third one at 10msec and so on). At what time will node D completely
receive the first three packets? And what will be the queuing delay for each packet (Justify
your answer using the following diagram.)

Pkt1

Pkt2

Pkt3

0 10 20 30 40

b) (2pt) What should be the maximum size of data packets such that there will be NO
queuing delay at any node? (Justify your answer. Here is another copy of the above
diagram, you might want to use it, but it is not necessary.)

Pkt1

Pkt2

Pkt3

0 10 20 30 40
Name:…………..
3
3. (3 points) Consider the sliding window protocols: Go-Back-N and Selective-Repeat that are
used to provide reliability. Consider the following four cases:
a. Sender uses Go-Back-N and receiver uses Selective-Repeat
b. Sender uses Go-Back-N and receiver uses Go-Back-N
c. Sender uses Selective-Repeat and receiver uses Selective-Repeat
d. Sender uses Selective-Repeat and receiver uses Go-Back-N
Compare the above four cases in terms of correctness of the protocol operation, i.e. whether the
reliability can be achieved or not. If reliability can be achieved, discuss how/why. If it cannot be
achieved, give an example case.

4
 Name:……………………..
4. (5 points) Consider the below network. Users X and Y are sending data to X’ and Y’,
respectively, using TCP Reno (the one that uses both timeout and 3 duplicate ACKs to
detect packet losses).
X 10 packets/sec X’
A B
Y Y’
Assume that our TCP implementation discards out-of-order segments. So, as in go-back-n, the
sender should re-transmit all packets starting from base when a packet loss is detected. Also, in
case of multiplicative decrease, X and Y always set Th to 4 (DO NOT USE Th= congwinw/2).

- Suppose at time t=0 the window size of X is 2, while that of Y is 6 as shown below,
- RTT is 1 sec for both connections X—X’ and Y—Y’, all the packets in the current window
are sent and ACKed in one RTT and the timers are set to one RTT.
- If router A gets more than 10 packets per second, it accepts 10 of them and drops the rest. But
while doing this, Router A gives higher priority to X-X’ connection (i.e., it first drops packets
from Y-Y’ connection).

a. (3pt) Give a detailed sketch of congestion window sizes at X and Y until at least t=20sec.
12
11
10
9
8
7
6
winX 5
4 Th
3
2
1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 sec

8
7
6
winY
5
Th
4
3
2
1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 sec

!!!!! Using these figures answer part (b) in the next page !!!!

5
Name:……………..

b. (2pt) What is the goodput for X and Y? Basically, find out how many actual application-
layer bytes are sent from X to X’ (or from Y to Y’) within [0, 20] sec. Then divide that
number by 20. Suppose that each TCP packet carries 1000-byte application-layer data.

6
Name:……………….
5. (4 points) Suppose an organization has the general network address of 175.95.67.128/25.
(! WHILE ANSERING THE FOLLOWINGS, JUSTIFY ALL YOUR ANSWER !).

a. (2pt) Now suppose that this organization wants to create 4 sub-networks

A with 30 hosts; B with 60 hosts; C with 6 hosts; D with 23 hosts;

Give a possible arrangement of sub-network addresses to make this possible.

b. (1pt) Based on your answer in part (b), determine how many more hosts can be
addressed in each network?

c. (1pt) Suppose a main router knows that the next node for A, B, C, and D are routers
Ra, Rb, Rc, and Rd, respectively. If that main router receives a packet with
destination address of 175.95.67.196, what would be the next node (again that
depends on your answer in part (b))?

7
Name:…………….

6. (2 points) Look at the BGP policy example given in slides Ch4 pages 109-110. Consider the
path information that reaches stub networks W, X, and Y. Based on the information available at
W and X, what are their respective views of the network topology. Justify your answer.

Hint: the following is the Y’s view of the topology.

legend: provider
network
X
W A customer
C network: