CITS3002 Computer Networks

Identify three distinct header fields that are required for the correct operation of both Data Link Layer and Network Layer protocols. For each of these three fields:

a.                      describe its typical role, and

b.                      provide an example to demonstrate why their 'duplication' does not introduce unnecessary redundancy.

2.                      [4 marks]        Consider a stop-and-wait protocol that employs only negative acknowledgments (NACKs), and no positive acknowledgments (ACKs).

a.                      describe what timeouts would need to be scheduled, and

b.                      explain why an ACK-based protocol is usually preferred to a NACK-based protocol.

3.                      [4 marks]        Nodes connected using standard wired Ethernet employ a digital encoding scheme that enables each node to detect frame collisions before each frame's checksum arrives.

a.                      explain how the standard's encoding scheme enables collisions to be detected.

b.                      Assume that Node-A and Node-B are connected by a single Ethernet segment. Node-A transmits a sequence of 7-bit ASCII 'A' characters and employs even parity to 'pad' the character to an 8-bit byte. Similarly, Node-B transmits a sequence of 7-bit ASCII 'B' characters and employs even parity to 'pad' each character to an 8-bit byte. (Unrealistically) assume that the transmitted bytes are not encapsulated with the standard Ethernet preamble, header, and checksum - only the characters 'A' and 'B' are transmitted.

DRAW AND DESCRIBE A LABELLED DIAGRAM which demonstrates how the transmissions of Node-A and Node-B are encoded on the wire, and clearly shows how the collision of the 'A-sequence' and the 'B-sequence' can be detected by each node.

4.                      [4 marks]        Consider the following network topology consisting of switches and hosts:

Hosts establish connections with each other using virtual-circuits before communicating any data. The following connections are established in the order given, and each connection is fully established before the next connection attempt is commenced. The virtual-circuit number assignment 'algorithm' always chooses the lowest unused virtual-circuit (added after test was released) number on each link, starting with zero. Assume that the sequence of connections is cumulative, that is, the first connection still exists when the second connection is established, and so on.

1.                      Host D connects to Host H.

2.                      Host B connects to Host G.

3.                      Host F connects to Host A.

4.                      Host H connects to Host C.

5.                      Host I connects to Host E.

6.                      Host H connects to Host J.

Give the virtual-circuit tables for each of the switches after all connections have been established.

5.                      [4 marks]        Consider the delivery of messages in an internetworked environment in which the source and destination nodes are many hops apart. Communication between distant nodes passes through a series of routers, and there may be multiple pathways between any source and destination. Large messages must be fragmented and reassembled using one of two possible strategies:

the first strategy involves fragmenting each message at the source node, and then reassembling them at the destination node.

the second strategy involves fragmenting each message at the source node, reassembling and re-fragmenting them at intermediate routers, and final reassembly at the destination node.

a.                      For each strategy, describe the fields of a Network Layer header necessary to support the strategy.

b.                      With reference to two distinct examples, describe circumstances where each of the strategies would be preferred over the other.