Data Communications, Spring 2000

2. Exercise (1. - 4.2.) (Tanenbaum, pp. 16-73)

1. Answer shortly the following questions ("review"-type of problem: answers are found directly in lecture notes or in Tanenbaum?s book):
a) What is the most important task of each OSI layer?
b) What are TCP and UDP? How do they differ?
c) How do concepts 'service', 'interface' and 'protocol' relate to each other?
d) What is meant by circuit switching, virtual circuit switching and packet switching?
e) What are protocols X.21, X.25, X.29 used for?
f) Why does ATM use small fixed-length packets?
g) What are ISO, ANSI, ETSI, ITU-T, CCITT?
h) Who are specifying ISO-standards, TCP/IP-standards?

2.OSI layers can be either connected-oriented or connectionless independently of each other. Give examples of environments (application <=> network) where it would be quite possible even reasonable to let adjacent layers to alternate between connection oriented and connectionless (include only those layers where connection-oriented and connectionless has some meaning).

3. In each layer many kinds of disturbance can happen such as

How can these be noticed? How to recover from the disturbance? Consider cases where the transferred information is a) a technical rapport b) a voice telephone call c) a video.

4. Data communication now and then causes congestion in the network. How is congestion noticed. What effects does congestion have? How the network can recover from congestion? Think what kind of solutions could be used in different OSI layers. (Don't go into details, principal ideas for different layers are sufficient.)

5. Why was SMDS designed as a conectionless network and frame relay as a connection-oriented one? How does their flow control differ? Is congestion anyhow possible? If so, what happens then?

6. Why is it so difficult to measure the data transfer time from the sender to the receiver acros WAN?

At least three problems should be solved including problem 1.