CPS 214 Computer Networks & Distributed SystemsFall 1994Midterm Exam (October 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. In which of the OSI reference model layers would you place the following tasks:
   1. Counting the number of words in a file and sending the result to a process running on another machine.

   2. Performing forward error correction across one particular satellite link in a network; error correction is not needed over the other links.

   3. Selecting a path through the subnet when a virtual circuit is created.

   4. Defining that two consecutive light pulses is a 1, two consecutive absences of pulses is a 0, and a combination of pulse-no pulse is the start of a new frame.

   5. Retransmitting packets that are discarded by the network layer of a connectionless architecture.

   6. Transparent compression of data.

   7. Checkpointing the transmission of extremely large files, so that the entire file need not be resent, should one of the machines crash during the transfer.

2. Consider a selective repeat sliding window protocol that uses 4-bit sequence numbers. For each of the following combinations of send and receive window sizes, state whether the resulting protocol is ``correct'' (e.g., suppresses duplicates, doesn't deadlock, etc.).
   1. Sending window (10 frames), Receiver window (6 frames)

   2. Sending window (6 frames), Receiver window (10 frames)

   3. Sending window (9 frames), Receiver window (9 frames)

   4. Sending window (15 frames), Receiver window (1 frames)

3. How many distinct signal encodings are needed to achieve a data rate of 100Mbps over a noiseless channel that filters out all frequencies above 100kHz and below 1kHz?  

4. In practice, no channel is completely noiseless. How high of a signal-to-noise ratio would a channel need in order to use the encoding in Problem 3? Explain.

5. The theoretical analysis of the efficiency of the ALOHA protocols showed that one could double the efficiency by changing the protocol from unslotted to slotted. What was the crucial factor in the analysis that doubled the protocol's efficiency in the latter case?

6. Consider an error detecting coding that uses 2 parity bits, with one parity bit checking all even-numbered bits, the other checking all odd-numbered bits. Is this code any better than the standard 1-bit parity method? Explain.

7.   Spanning tree bridges have the drawback that they forward frames along a spanning tree, even though a shorter path might exist between the two communicating hosts. What is the reason for this restriction?

8. If you were designing a generator polynomial for a CRC checksum function, which of the following would be a better choice? Hint: Which generator detects more errors? Explain.
   1. , where has degree 16.
   2. , where has only degree 8.

9. At the physical layer, FDDI encodes bits using a code referred to as ``4B/5B.'' In 4B/5B, every 4 data bits are transmitted as 5 ``cells'' (similar to a baud), with each cell consisting of a pulse or absence of light. The extra cell is used for clock synchronization. Does 4B/5B encoding have any advantages over Manchester Encoding?

10. Some people argue that the OSI reference model is inherently flawed. Give 2 reasons that are commonly cited.

11. Consider the implementation of a sliding window protocol across a 100Mbps fiber link (one way propagation delay is 6 ms). If the sequence number space limits the size of the send window to 255 frames, how large should frames be in order to achieve a channel efficiency of 100%? You may assume that ACKs are piggybacked on returning data frames.

12. The old Arpanet routing algorithm suffered from the problem ``good news travels quickly, bad news travels slowly''. Describe the new Arpanet routing algorithm, explaining how this problem was solved.

13. Consider a phone connection between someone in Durham and India (e.g., halfway around the world). BT&'s long distance service uses digital transmission to produce clear connections. In contrast, FLEECE (an alternate long distance service provider) uses totally analog circuits, resulting in connections in which both parties can barely hear each other. If the signals in each case travel the same distance and traverse the same number of switching nodes, what accounts for the difference in quality? In particular, what happens to the signals as they are processed by intermediate switching nodes?

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 mid.tex.

The translation was initiated by Thomas Narten on Sat Oct 12 13:47:46 EDT 1996