CPS 214 Computer Networks & Distributed SystemsFall 1994Final Exam (December 12)Prof. Thomas Narten

NAME: SCORE:

This is a closed book (and notes) examination; however, you may consult your index card. Answer all questions on the exam itself. Justify your answers! You must explain your reasoning to receive full credit on yes/no type questions.

All questions have equal weight. However, some questions are harder than others; budget your time accordingly. You may omit one question, but you must clearly mark which question should not be graded, or I will assume that you wish to omit the final question on the exam.

1. What services does UDP provide above and beyond what IP already provide?

2. Suppose hosts A and B are connected by a single point-to-point physical link that uses a sliding window protocol (window size of 4 frames) with acknowledgments and retransmissions at the data link layer. In addition, A and B have agreed to uses a maximum segment size (MSS) of 512 bytes. If A starts with a send window size of 1024 bytes and doubles it to 2048 bytes, is throughput likely to increase? Explain.

   What is the likely effect of further doubling the TCP window size to 4096 bytes?

3. SIP addresses are 64 bits long, twice as long as IP addresses. Give at least two reasons why larger addresses were chosen.

4. Routing protocols such as RIP allow a gateway to select a path through the neighboring gateway having the lowest cost to a given destination. What additional information does BGP provide, and how would a gateway use that information?

5. The SIP scoping mechanism for multicast packets makes scope an explicit part of a multicast address. In contrast, IP multicasting simply uses IP's time-to-live field. What advantage does SIP's mechanism have over IP's?

6. Is it possible to further fragment an already fragmented IP datagram? If so, explain how, if not, explain why it can't be done.

7. All ICMP messages include the complete IP header and part of the data portion of the IP packet that caused the error. In the specific case of ICMP redirects, is the IP header information actually needed in order to correctly update the sending host's routing tables?

8. In the implementation of RPC discussed in class, the client presents the server with a ``unique identifier'' and ``table index'' whenever it makes an RPC call. What field in the TCP header does the ``table index'' most closely correspond to? Explain.

9. The TCP push mechanism is intended to allow the sending application to inform TCP that it should send any buffered data. Since this can be achieved through the interface between the application and TCP, why does TCP include the push bit in its header?

10. Give an example that demonstrates why the TCP urgent mechanism is needed.

11. Is it ever useful for TCP to advertise a flow control window of 0 bytes? Explain.

12. Herb Gritz is studying the performance of TCP. In his experiments, he increases the size of the send and receive windows in order to find a value that maximizes end-to-end throughout. At first, increasing the size of the window does improve performance. Then, much to his dismay, throughput drops suddenly. Indeed, he observes good performance for a window size of X, but noticeably worse performance for a size of W+1. What accounts for the observed behavior?

13. Would it be possible to use the DNS to map IP addresses into their corresponding physical addresses (e.g., as a replacement for ARP)? What would be some of the advantages and disadvantages of doing so?

14. Consider the authoritative section of DNS responses. Is this field meaningful only of the server doesn't know the answer to a query? Explain.

15. Consider an RPC implementation in which the server is stateless. Is it possible for the RPC system to provide ``at most once'' semantics? What about ``exactly once'' semantics?

16. Is DES encryption vulnerable to statistical attacks? If not, why not? Otherwise, explain why it is not a concern.

17. Is it possible to prevent message tampering (e.g., unauthorized modification) without encrypting the entire message? Explain.

18. SMTP includes a command ``MAIL FROM: <reverse-path>'' that the client sends to the server. Why is this command present, when it is also the case that the sender's return address is contained in one of the header lines that is sent as normal data sent as part of the ``DATA'' command?

19. What is an HTML anchor and for what is it used?

20. Mobile computing deals with allowing a machine (e.g., a laptop) to move its physical location frequently, yet have access to network services whenever it is connected to the Internet (regardless of the machine's physical location). What is the primary issue that makes mobility difficult to solve?

About this document ...

This document was generated using the LaTeX2HTML translator Version 0.6.4 (Tues Aug 30 1994) Copyright © 1993, 1994, Nikos Drakos, Computer Based Learning Unit, University of Leeds.

The command line arguments were:
latex2html -split 0 final.tex.

The translation was initiated by Thomas Narten on Sat Oct 12 13:52:55 EDT 1996