Exercise 8 (14.-17.3.) (Tanenbaum pp. 304-374)

1. Answer shortly the following questions.
a) What is FDDI?
b) How high data rates and how long distances are possible for the 802.3u standard (Fast Ethernet)?
How is it possible to detect collisions?
c) What is HIPPI?
d) What is the task of the network layer?
e) What can be done with the Bellman-Ford algorithm?
f) Why, in distance vector routing, do good news spread fast and bad news slowly?
g) What does "split horizon" mean?

2. A set of LANs are connected through bridges ( see the figure below).
a) Give an example of a situation where the routing tables for the bridges cannot be formed by mere flooding.
b) Construct the spanning tree (bridge B1 as the root node) for the routing.
c) Simulate the transmission of a frame from a node in L3 to a node in L1. Why is the transparent bridge not always
making optimal usage of the bandwidth?

   
LAN L1  XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX                   
           |                     |                     |                   
           B1                    |                     |                     
           |                     B6                    B5 
LAN L2  XXXXXXXXXXXXXXXXXXX      |                     |                    
           |      |              |                     |                        
           |      B3             |                     |                    
           |      |               \                    |                  
          B2    XXXXXXXXXXXXXXXXXX|XXXXXXXXXXXXXXXXXXXXXXXXXXX  LAN L3                   
           |                      /             |                   
           |                     |              B4                    
           |                     |              | 
LAN L4  XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX   

3. Suppose that bridges in the picture are source routing bridges. In the beginning routing tables are empty. Then a node in the LAN L3 decides to send a message to a node in the LAN L1. Explain, in reasonable detail, how the sending of the message is conducted. Continue until the benefits and drawbacks of source routing have clearly been exposed.

4. a) A large FDDI ring has 100 stations and a token rotation time of 40 msec. The token holding time is 10 msec. What is the maximum achievable efficiency of the ring? How big is the overhead?
b) Consider building a data channel for a supercomputer using the HIPPI approach , but modern technology. Each word has now 64 bits and a word can be sent every 10 nsec. What is the data rate of the channel?

5. a) Give a simple heuristic algorithm for finding a 'fault tolerant' path from a network node A to another network node B. Fault tolerance here means that the system survives the loss of one link. The routers are considered reliable enough, so it is not necessary to worry about the possibility of router crashes.
b)One problem in flooding is that packets remain circulating in network loops forever. What different ways there are to settle this problem?

6. a) What kind of distance metrics can be used in routing tables? What are the benefits and drawbacks of these metrics?
b) How do distance vector routing and link state routing differ from each other? Consider all the stages of link state routing.