TDC 375                                                 John Kristoff
Mid-term Exam Guide                                 February 10, 2002


Exam Format

The midterm exam will consist of multiple choice, fill in the blank and
short essay questions.  The exam will take you about 1 to 1 1/2 hours to
complete, but you have as much time as you need.  If you are having an
unreasonably difficult time with the exam, chances are others are as
well.  Do the questions you know best first and go back to the hard ones
later.  There is a grading curve for the final grade to correct any
problems in the exam or my instruction that led up to the exam.

The exam is closed book, closed notes.  No calculators, computers or
electronic devices are allowed.  Only basic math skills are necessary
for the exam.  You will have enough paper as necessary, but if you need
more you can ask the proctor for additional sheets.  A final word of
caution, DO NOT CHEAT!

Topics Covered

The book, required reading and optional reading should be used as
reference material to study from.  However, class lectures and class
discussions are as equally as important references for the exam.  Below
is a list of concepts you should have mastered in order to do well on
the exam.

I. Introduction to Network Protocols
   A. Addressing
   B. Fragmentation
   C. Routing
   D. Internet Protocols
   E. The end-to-end argument
   F. IETF
   G. NAPs and Exchanges

II. Internet Protocol (IP)
    A. Addressing
    B. IPv4 addressing issues and solutions
    C. IPv6

III. Network Control
     A. ARP
     B. DHCP
     C. ICMP

IV. Routing Protocols
    A. Static
    B. RIP
    C. OSPF
    D. BGP

V. IP Multicast
   A. Addressing
   B. IP to Ethernet MAC mapping
   C. IGMP
   D. DVMRP
   E. PIM-DM
   F. PIM-SM
   G. MBGP
   H. MSDP

In addition to the topics above, you should be familiar with the tools,
discussions and examples we used in class.  Please review your course
notes, coure online material and course homepage for reference.

Sample Exam Questions

1. What is the difference between unicast, multicast and broadcast?

2. What will a router do with an IP packet whose TTL field is set
   to zero?

3. How many hosts can be uniquely addressed if the subnet mask is
    255.255.252.0?

4. Would RIP or OSPF give better convergence time?

5. If a station's gateway is 10.255.10.254, what is its subnet
    mask? (assume classless addressing)

    a. This is an illegal address
    b. 10.255.255.255
    c. Cannot be determined
    d. This is a loopback address

6. What is the difference between subnetting and supernetting?

7. What are the documents used by the IETF to document practices
    and standards for the Internet?

8. What field in the IP header protects against routing loops?

9. In the original class hierarchy of IP addressing, which
   class of addresses would an address fall if its first 4 bits
   started with 1100?

10. What is the subnet mask /20 in dotted decimal notation?

11. Compare and contrast what you know about the end-to-end argument
    to what you know about the telephone network.  How is or how isn't
    the telephone network end-to-end?

12. What part of the IP datagram makes the 'traceroute' utility
    possible?

13. What is the primary capability that IGMP version 3 offers over
    previous versions of IGMP?

14. Give four common pieces of information that a DHCP server will
    give to a DHCP client.

15. What problems with the Internet does Steve Deering see today?
    What is one of his recommended solutions as described in the
    hourglass presentation?

16. What is the purpose and benefit of a network access point (NAP)
    for Internet connectivity?  Who uses a NAP?

17. Describe the process of discovering an IP host's Ethernt MAC
    address.

18. Discuss IP fragmentation.  How is it done?  Why is it done?
    What is good or bad about fragmentation?

19. What was the apparent problem to Internet routing table growth
    as presented by Randy Bush?  What was one of the solutions as
    described in his presentation?

20. When would you want to use flood and prune multicast as opposed
    to explicit join multicast protocols?

21. Marshall Eubanks discussed single source multicast (SSM), why do
    broadcasters prefer SSM as he described it in his presentation?

22. You start up your PC and load a local HTML file that contains a
    link to http://www.yahoo.com.  Assume no local web caching is
    being done.  Describe the network communications that takes place
    to pull up Yahoo!'s web page.

23. Firewalls, NATs and Routers.  Describe how each of these boxes
    are in opposition to the end-to-end argument or not.

24. ARP is inherently insecure.  Discuss what sorts of problems an
    attacker could use by abusing ARP messages.

25. Why does using a rendezvous point (RP) cause scaling problems in
    IP multicast?  Are there any solutions to this scaling problem?

26. How does RIPv2 different from RIPv1?

27. True or false, link state routing protocols are better are
    preventing routing loops than distance vector routing protocols.

28. True or false, distance vector routing protocols converge faster
    than link state routing protocols.

29. There are a number of potential interesting uses and abuses with
    ICMP as described in the class lecture and in Ofir Orkin's
    materials.  Describe how it is possible to determine a host
    operating system simply by the use of ICMP scanning techniques.

30. The end-to-end argument would say that 'network state' is something
    to avoid.  What is network state?

[end]