Net-Centric Computing, 2017
Midterm
Midterm Examination 2017
(Open books, slides)
Time Allowed: 75 minutes (Wednesday, March 15, 2017)
Name
------------------------------------Student ID ------------------------------------Instructions:
 Please write your name on the exam
 Write something for every question, anything outside the box will NOT be graded.
 You will get points if you attempt a solution but nothing for a blank sheet of paper
 Problems take long to read but can be answered concisely
 Anyone looking at a neighbors paper will be told to leave
 There will be no sharing of books/notes
 Books and slides are allowed, anything else is not.
 Show all your work. Partial credit is possible for an answer, but only if you show the
intermediate steps in obtaining the answer.
 Good luck.
Problem
1
2
3
4
5
6
7
8
9
10
11
12
Marks
5
5
5
5
5
10
10
10
10
10
10
10
Total
125
d
1. Consider the Alternating-Bit protocol, assume that the receiver is expecting a segment with sequence
number 0, but instead the receiver receives a segment with sequence number 1. What response does the
receiver give??(5 marks)
2. What is the purpose of using a randomly selected initial sequence numbers in TCP, why not start at zero
each time? (5 marks)
3. What is the purpose of using a randomly selected string as a cookie, What if its not random? (5 marks)
4. Why would an organization get multiple consecutive class C addresses instead of randomly assigned
addresses (5 marks)
5. Caching is a commonly used technique to reduce traffic and latency. However, one potential drawback of
caching is that cached data may be stale. What feature of HTTP allows a browser to guarantee the freshness
of a web page when it uses its browser cache? (5 marks)
Net-Centric Computing, 2017
Midterm
6. Outline the design of a DHT based ‘discussion’ service that allows you to anonymously discuss net-centric
assignments with your classfellows without your TAs being able to catch you?(10 marks)
7. Ahsan decides to start a small company. He asks his ISP, IST Networks, to give him enough addresses for
1500 hosts. IST allocates a subblock from the 192.1.* address range that they own and tells Ahsan to use the
following addresses:
192.1.0.*
192.1.1.*
192.1.2.*
192.1.3.*
192.1.4.*
192.1.5.*
a) Ahsan has heard that the size of the Internet routing table has grown to huge proportions, and that to be a
good citizen, he should announce the fewest number of routes possible to exactly cover his IP addresses.
Under CIDR, what is the smallest set of network numbers that the rest of the world would use to describe
Ahsan’s networks (please use address & prefix format)? (5 marks)
a) Ahsan has a second ISP that he uses, PTCL. He announces some of his network addresses to both IST and
PTCL. (Ahsan hadn’t talked to you yet—these announcements do not correspond to the answers you
provided to the previous question). As a result, some router far away in the network produces a forwarding
table with the following entries:
Destination
192.1/16 1.2.3.4
192.1.4/23 1.2.3.5
192.1.1/24 1.2.3.6
192.1.0/23 1.2.3.7
Which next hop should the router use for a packet destined to 192.1.0.1? (5 marks)
8. Nofel is working on a Web site that has multiple replicated servers located throughout the Internet. He plans
on using DNS to help direct clients to their nearest server, and comes up with the following hierarchical
scheme. Nofel has divided his server replicas into three groups (east, west, and central) based on their
physical location. A typical query occurs as follows:
 When a client makes a query for www.cheetay.ta.com, the root .com name server is contacted first.
It returns the name server (NS) record for ns1.ta.com (along with a corresponding A record). The
TTL of this record is set to 1 day.
 The ns1.ta.com name server is then queried for the address. It examines the source of the name
query and returns an NS record for one of {east-ns, central-ns, west-ns}.cheetay.ta.com (along with
a corresponding A record). The choice of which name server is based on where ns1 thinks the query
came from.
 Finally, one of {east-ns, central-ns, west-ns}.cheetay.ta.com is contacted and it returns an address
(A) record for the most lightly loaded web server in its region.
The following questions are based on this design.
Nofel’s name server software has only two choices for TTL settings for A and NS records - 1 day and 1 minute.
What are reasonable TTLs for the following records? Briefly explain your choice.
a) NS record for {east-ns, central-ns, west-ns}.cheetay.ta.com:
b) A record for {east-ns, central-ns, west-ns}.cheetay.ta.com:
c) A record returned for the actual Web server: (10 marks)
9. Outline/Sketch an algorithm or write pseudo-code that would allow you to determine the number of active
web-servers in the internet. (10 marks)
10. Consider the following plot of TCP window size as a function of time. Answer the following questions and
provide a short discussion justifying your answer.
Net-Centric Computing, 2017
Midterm
a.
b.
c.
d.
e.
f.
g.
h.
a)
b)
c)
d)
e)
f)
g)
h)
identify the interval(s) of time when TCP slow-start is operating
Identify the interval(s) of time when TCP congestion avoidance is operating
After the 16th transmission round, is segment loss detected by a triple duplicate ACK or a timeout
After the 22nd transmission round, is segment loss detected by a triple duplicate ACK or a timeout
What is the initial value of the threshold at the first transmission round
What is the value of the threshold at the 18th transmission round
During what transmission round is the 70th segment sent
Assuming a packet loss is detected after the 26th round by the receipt of a triple duplicate ACK.
What will be the values of the congestion window and the threshold.
11. You are asked to design the ticketing system for a new bus service (don’t worry we bought our system in
Turkey ). When you ride the bus, the fare system needs to track the value of your ticket along with where
you enter and where you exit the bus system in order to compute and deduct the fare to charge you. To do
so, it uses information encoded in your ticket.
Consider yourself as a “client” of the fare system and the ticket machines and the entry/exit gates as
interfaces to the fare system’s “server.” You connect to the server when you buy a ticket, and again each
time as you enter the subway and later exit it.
(a) If the fare system stored per-client state in the server, what minimal information would need to be encoded in
your ticket?(2 marks)
b) If the fare system used stateless operation, what information needs to be encoded in the ticket?(2 marks)
c) Which of these corresponds to HTTP’s operation, and what is the HTTP component that the information on
the ticket corresponds to?(3 marks)
d) The Bus system in fact uses stateless operation. Give two advantages and disadvantages of this.(3 marks)