Question 1
1.a The network core is a mesh of interconnected routers used to deliver packets worldwide.
i. Explain the purpose of the Store and Forward technique in packet switching. [2 marks]
ii. In the context of packet switching, Loss and Queuing Delay can occur. Explain the relationship between the two, being specific on when one may lead to another. [3 marks]
iii. Compare and contrast the principles of Forwarding and Routing, with specific to reference on how they are often used together. [3 marks]
[8 marks]
1.b Access networks are a key part of the network edge, with different technologies enabling users to connect to the wider Internet.
i. State and briefly describe two of the technologies that are used in residential (home) settings. [3 marks]
ii. Compare the two technologies identified, specifically commenting on how the underlying access medium is shared between multiple users and also other concurrent services. [3 marks]
[6 marks]
1.b Throughput is a measurement of the rate at which bits can be transferred between a se nder and receiver.
i. When measuring per-connection end-to-end throughput, state how the bottleneck c an be determined, given the available bandwidth of each individual link. (1 mark)
ii. Consider the network topology below. This contains three hosts (H1, H2 and H3) connected through two network nodes (N1 and N2). All network links have equal throughput. Hosts H1 and H2 send enough packets to saturate their initial links to N1. Applying your knowledge of Packet Loss and Throughput, explain what will happen in the network. Consider also the actions that the H1 and H2 may take to mitigate any issues. (6 marks)
iii. Name and briefly describe two techniques used by network nodes to control Packet Loss. (4 marks)
[11 marks]
Total 25 marks
Question 2
2.a NAT is an approach used to tackle a shortage in global IP addresses.
i. Explain the functionality of a NAT router, making reference to the operations applied to both outgoing and incoming datagrams (4 marks)
ii. You have been tasked with creating a campus LAN for a university. This must support a large number of connected hosts, including university web servers. Applying your knowledge of NAT, explain whether a single-address NAT router is suited to the application. Justify your answer and include any alternative solutions that could be used instead. (5 marks)
[9 marks]
2.b DHCP is used to manage the network configuration of devices when joining a network.
i. Describe the interaction between a host and DHCP server when a host joins a new network. Assume that the host’s networking device is unconfigured. Illustrate this interaction through a flow diagram, demonstrating clearly the purpose and DHCP message type used at each stage. (8 marks)
ii. State two different types of leases that a DHCP server may distribute to a client, briefly describing each. (4 marks)
iii. Using your knowledge of the various network layers, describe each step of the process that would be undertaken to encapsulate a DHCP packet and send it from the host. (4 marks)
[16 marks]
Total 25 marks
Question 3
a. Assume you have been given the 10.1.0.0/24 IP prefix. You need to subnet the following network topology.
i. How many subnets should there be? [3 marks]
ii. Each subnet should support the same number or hosts. What is the subnet mask (in dotted decimal format) to achieve this addressing scheme? [3 marks]
iii. How many usable IP addresses (addresses that can be allocated to a host) does each subnet support? [3 marks]
iv. Is there any subnet sub-prefix available for future use? [2 marks]
v. Based on your addressing scheme, give some sample addresses for • Server 1 • Server 2 • Server 3 • Router interface iface 3/1 that connects Router 3 to Router 2 [4 marks]
Explain each of your answers. [15 marks]
b. Assume you are a network administrator managing a network with the topology shown in the following figure.
To load balance the traffic to the two subnets, you configure Router 1 to export the IP prefix 90.252.100.0/24 only to Router 2 and export the IP prefix 90.252.101.0/24 only to router 3, so that the links 1-2 and 1-3 carry half of the traffic each. However, you notice that when a link fails the corresponding subnet becomes unreachable, i.e. if Link 1-3 fails the subnet 90.252.101.0/24 will be unreachable.
What additional prefix advertisement you can make from Router 1 to routers 2 and 3 so that you maintain the same load balancing features, but in case of a link failure the other link can be operate as a fallback link. Namely, when both links work traffic to 90.252.100.0/24 should still go through Link 1-2, while traffic to 90.252.101.0/24 should still go through Link 1-3. But if either link fails the other link should be able to route the traffic to both subnets. [5 marks]
c. You are a website administrator, and you operate the website for a Lancaster venue. The website runs on a server that is located in Lancaster, and traffic to this website is also predominantly from the local area. However, the server has connection has limited bandwidth and at certain times of the day it can get congested. Someone suggests that hosting the website in a well-provisioned CDN will solve the issue, but the nearest CDN nodes are in London while the CDN operates multiple other nodes around Europe. What would be the advantages and disadvantages of this solution in terms of performance and resilience? [5 marks]
Total 25 marks
Question 4
a. You type the following URL into your web browser
Assuming that
• The address of the DNS resolver is 1.1.1.1
• Neither your host nor your DNS resolver have any cached DNS entries
• DNS never needs to fail over to TCP
• The HTML response returns 200 OK with a web page
• The HTML request and response each fit in a single segment
• The web page requires loading no additional resources
Write down the series of packets exchanged that will occur for your host to receive the web page. Include packets sent by your DNS server, and control packets for TCP connection setup and teardown. You need not include any ARP packets, and you do not need to write down message formats. Simple descriptions such as “X sends a UDP segment to the HTML server on the HTTP port Y” are sufficient. In the case of the HTTP request, clearly state the path of the file requested in the GET. [12 marks]
b. The diagram below shows two subnets connected by a router. For each host and router port, the IP address and MAC address (abbreviated) are shown. Initially the ARP tables of the two hosts and router are empty. Suppose A sends a packet to B
i. Write the content of the ARP tables for the two hosts and the router after the packet reaches host B. [4 marks]
ii. Suppose that in the subnet in which Host A is connected, another 100 hosts are also connected. The hosts do not communicate with each other but they all exchange traffic with the server. How many entries would be in the ARP tables of Host A, Host B and the router? Explain why. [3 marks]
[7 marks]
c. Consider the topology below as part of the Internet map. Circles represent Autonomous Systems (ASes) that connect to other networks using the Border Gateway Protocol (BGP). A single-ended arrow indicates a “customer-to-provider” relationship, while dashed lines indicate “peer-to-peer” relationships. For example, network B is a customer of ASes 5 and 2, while AS1 and AS2 are peers. Paths are selected at each AS based on the normal customer/provider/peer policies, and no other policy is applied.
AS 5 received some reports from its users that they cannot access every other AS in graph. You are a renowned BGP expert and they ask you to help them resolve this issue.
i) Which ASes are not reachable from AS 5 and why? [3 marks]
ii) How can AS 5 change its existing relationships to make sure that it can get routes to every other AS? Explain why. [3 marks] [
6 marks]
Total 25 marks
Comments