CPS 214 Computer Networks & Distributed Systems Fall 1996
Midterm Exam (October 16) Prof. Thomas Narten

NAME:$\underline{\hspace*{2in}}$ SCORE:$\underline{\hspace*{2in}}$

This is a closed book (and notes) examination; however, you may consult your $3\times 5$ 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.

Answer true or false for each of the following assertions:

(a)

The Internet is growing geometrically in size.

(b)

On broadcast LANs, it is trivial to support efficient multicasting.

(c)

The ISOC/IETF is generally viewed more positively than the ISO as a standards-setting organization.

(d)

ATM carries fixed-sized 32-byte cells.

(e)

Ethernets perform very poorly under heavy loads due to excessive collisions.

(f)

Stations on an IEEE 802.5 token ring add a 24-bit delay to each frame they forward; the number 24 comes from the number of bits in the token.

2.

Suppose that a client requests an n-byte data file from a server. Which has greater impact on the total speed of the transfer, network latency or network bandwidth? Explain.

3.

When speaking of network layering, the terms network service, service interface and protocol often come up. What do they mean and how do they differ from one another?

4.

In observing the process of standards development, Dave Clark made an observation called the ``apocalypse of the two elephants''. What did he mean?

5.

At the physical layer, one frequently talks about the modulation rate (i.e., baud rate) and the encoding method. What do these terms mean and how do they relate to a link's data rate?

6.

Is it possible to design an error detecting code that catches (i.e., detects) all errors for a packet whose size is bounded by a known maximum value? Explain.

7.

Today's modems achieve 28.8kbps throughput over voice grade phone lines. Are further significant improvements over such lines likely in the future? Explain.

8.

RS-232-C does not use Manchester encoding. Explain why it is not needed, or how RS-232-C addresses the problem solved by Manchester encoding.

9.

One of the jobs of the Data Link layer is to identify the start and end of frames. Given a choice, would you use encoding violations or sentinel values to recognize frame boundaries? Why?

10.

The Ethernet design doesn't scale well to WANs because the increased latencies cause the cost of collisions to become excessive. Likewise, the IEEE 802.5 Token Ring design would lead to poor throughput if the ring latency increased substantially. What changes to the basic token ring design are needed to give it reasonable throughput on rings with high latencies (e.g., FDDI)?

11.

The exponential backoff algorithm used by the Ethernet has colliding stations wait a (random) multiple of a slot time (i.e., 51.2$\mu s$) before attempting transmission. Why don't stations simply delay a random amount of time taken from a uniform interval? Be specific!

12.

The Revised Arpanet Routing Metric was developed in response to problems with using delay as a metric. What are the problems associated with using delay as a routing metric? (Note: I'm looking for problems that are inherent to using delay as a routing metric in the general case, rather than problems related to the specific Arpanet implementation.)

13.

Describe the link-state routing algorithm used in the Arpanet and explain why permanent routing loops cannot form.

14.

Give three examples of how the x-kernel helps protocol developers construct efficient protocol implementations.

15.

Consider a sliding window protocol operating at the data link layer.

(a)

What is the minimum window size need in order to achieve 100% utilization of the link?

(b)

What is the minimum number of bits in the sequence number required for correct operation of the protocol?

16.

Consider 4 users sharing a network link. In each of the following, explain whether time-division multiplexing (TDM) or Statistical TDM (if either) would be more appropriate:

(a)

Each user is generating data at a steady rate via pulse code modulation (PCM).

(b)

The average rate of each of the four senders is the same, but traffic from each source is bursty.