Overcoming Traffic Congestion in Ethernet Networks

This document provides a detailed analysis of various networking concepts, starting with Ethernet networks and progressing to IPv4 and IPv6 protocols, and finally, routing protocols. In the first section, it addresses the challenges of managing traffic in modern switched Ethernet networks where routers might receive more traffic than they can forward. It explores how polling can mitigate this issue, outlining its advantages and disadvantages. The discussion then shifts to Token Bus networks, explaining their transmission mechanism and how their performance changes with increasing transmission speeds. The impact of collisions and broadcasts on Ethernet network throughput is analyzed, along with the roles of repeaters, hubs, bridges, and switches in influencing these factors. The second section focuses on IPv4, examining why only the header is checked for errors despite Ethernet's CRC check for the entire frame. It demonstrates the calculation of a valid IP address range given a subnet mask. The fragmentation mechanism in IPv4 and its contribution to network congestion are explained. A negative acknowledgment mechanism for IPv4 is proposed, detailing its operation through a time diagram and discussing its advantages and disadvantages. The third section compares RIPv2 and OSPFv2 routing protocols, clarifying the nature of the information they receive from neighbors and why it's considered "second-hand" or "first-hand." It demonstrates the Shortest Path First algorithm using a sample Link State Database and network topology. Finally, it compares and contrasts key fields in IPv4 and IPv6 headers, explaining their purposes, relationships, and the reasons behind any changes made between the two versions. The document provides a thorough understanding of these networking concepts, suitable for students and professionals in the field.

Question 1 (7 + 8 + 10 = 25 marks)
a) In modern switched Ethernet networks, it is not unusual for a router to receive more traffic
than it is able to forward due to the relative speeds of the Ethernet network versus the
links from the router to other networks or the Internet. In such circumstances, packets
would be randomly dropped either by the router or by the interconnecting switch when
buffers are filled. Briefly discuss how polling could be used to overcome this problem, and
discuss two advantages and two disadvantages of this approach.
b) Prior to the development of full-duplex Ethernet, Token Bus networks provided a means
through which communication could be achieved on a shared medium without collisions.
Briefly describe how devices transmitted on a network using Token Bus (ignore token
management) and discuss the performance token bus as transmission speed increases.
c) Collisions and broadcasts have a negative impact on the throughput of an Ethernet
network.
i) Describe how collisions and broadcasts influence the throughput on an Ethernet
network using a bus topology.
ii) Repeaters, hubs, bridges, and switches are all devices that can be used in the
construction of Ethernet networks. Briefly discuss how each of these devices
would influence the impact of collisions and broadcasts within a single Ethernet
network. Hint: VLANs divide a network into two or more networks and is irrelevant
to this question.
Question 2 (6 + 6 + 8 + 20 = 40 marks)
a) IPv4 uses a checksum to determine if any errors have occurred in the transmission of a
datagram, however this checksum only checks for errors in the IPv4 header. Discuss why
only the header is checked by IPv4 for errors, and discuss whether it is necessary to check
for errors at all given that Ethernet already checks the entire frame using CRC.
b) Given a subnet mask of 255.255.192.0, demonstrate for an IP address of your choosing the
calculation of the range of addresses that are valid from that network (host ID all 0s
through to host ID of all 1s).
Note: The IP address you choose must consist of four decimal numbers that are greater
than or equal to 101, and must be odd numbers.
c) The fragmentation mechanism provided by IPv4 can contribute to network congestion.
Briefly describe how fragmentation works and, given that there is no change to higherlevel
data, explain how this mechanism could contribute to congestion.
d) Consider implementing a negative acknowledgement mechanism for IPv4 that would alert
a source host to a datagram being discarded due to a corruption in the path between the
communicating hosts at any router in the path or the destination host.
i. Briefly describe how your protocol would work. ii. Illustrate the operation of
your protocol using a time diagram/sequence diagram (see Session 5 Slide 39 for an
example), including both successful and unsuccessful delivery of datagrams.
iii. Discuss two advantages and two disadvantages of your protocol.
Question 3 (11 + 8 + 16 = 35 marks)

a) Both RIPv2 and OSPFv2 routing protocols receive updated information about the state of
the network from their neighbours. However it can be said that the decisions made by
vector-distance routing protocols (such as RIPV2) are based on “second hand” information,
whilst the decisions made by link state routing protocols (OSPFv2) is based on “first hand”
information. Briefly describe what information is contained in the updated information
routers receive, where the information comes from, and why it is considered “second
hand” or “first hand”.
b) Given the following Link State Database and network topology:
A B C D E F
A - - 1 - 4 1
B - - - 3 1 6
C 1 - - - 2 -
D - 3 - - - 2
E 4 1 2 - - -
F 1 6 - 2 - -
Demonstrate the working of the Shortest Path First algorithm and final routing table for
one of the routers of your choosing. Your answer must indicate the paths and costs that
will be taken by all datagrams through the network.
c) For each of the fields listed below appearing in the IPv4 and IPv6 headers, briefly describe
the purpose of those fields, explain how they are related, and explain why a change has
been made:
i. IPv4 defines both header length (HLEN) and total length fields, whilst
IPv6 defines only a payload length.
ii. IPv4 defines a protocol field, whilst IPv6 defines a next header field. iii.
IPv4 defines a time to live (TTL) field, whilst IPv6 defines a hop
limit field.
Marking Scheme
A
C
E F
B
D

Question 1 (7 + 8 + 10 = 25 marks)
• Part (a) o (3 marks) Discussion of how polling could be used.
o (2 marks) Discussion of advantages.
o (2 marks) Discussion of disadvantages
• Part (b) o (3 marks) Description of transmission using token
bus.
o (5 marks) Discussion of performance of token bus.
• Part (c)(i) o (2 marks) Description of how collisions influence
throughput.
o (2 marks) Description of how broadcasts influence throughput.
• Part (c)(ii) o (3 marks) Discussion of influence of repeaters
and hubs.
o (3 marks) Discussion of influence of bridges and switches.
Question 2 (6 + 6 + 8 + 20 = 40 marks)
• Part (a) o (3 marks) Discussion of why only the header
checked.
o (3 marks) Discussion of why check needed given Ethernet CRC.
• Part (b) o (2 marks) Correct conversion of IP address and
subnet mask to binary. o (2 marks) Correct application of
binary AND.
o (2 marks) Correct identification of first and last addresses.
• Part (c) o (3 marks) Description of how fragmentation works.
o (5 marks) Discussion of how fragmentation can contribute to congestion.
• Part (d) o (3 marks) Description of how protocol works. o (2
marks) Illustration shows successful delivery of datagram. o
(3 marks) Illustration shows unsuccessful delivery of
datagram. o (6 marks) Discussion of two advantages.
o (6 marks) Discussion of two disadvantages.
Question 3 (11 + 8 + 16 = 35 marks)
• Part (a) o (3 marks) Explanation of information contained in
RIPv2 updates. o (3 marks) Explanation of information
contained in OSPFv2 updates. o (2 marks) Correctly identifies
source of information.
o (3 marks) Explanation of why information is second/first hand.
• Part (b) o (5 marks) Correctly indicates paths and cost for all
destinations.
o (3 marks) Correct working.

• Part (c)(i) o (3 marks) Description of fields.
o (3 marks) Explanation of relationship and changes.
• Part (c)(ii) o (2 marks) Description of fields.
o (3 marks) Explanation of relationship and changes.
• Part (c)(ii) o (2 marks) Description of fields. o (3 marks)
Explanation of relationship and changes.