Primary Examination, Semester 1, 2016 Official Reading Time: Writing Time: Total Duration: Questions Answer all 5 questions 10 mins 120 mins 130 mins Computer Networks & Applications COMPSCI 3001, 7039 Instructions • Begin each answer on a new page in the answer book. • Examination material must not be removed from the examination room. Materials • Simple calculators without alpha-numeric memory or communication capabilities allowed. • Foreign language paper dictionaries permitted. DO NOT COMMENCE WRITING UNTIL INSTRUCTED TO DO SO Time 120 mins Marks 120 marks 120 Total COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 2 of 9 Application Layer Question 1 (a) What information is used by a process running on one host to identify a process running on another host? (b) Why does HTTP run on top of TCP rather than on UDP? (c) Consider an HTTP client that wants to retrieve a Web document at a given URL. The IP address of the HTTP server is initially unknown. What transport and application-layer protocols besides HTTP are needed in this scenario (list all that will be used)? (d) Consider a short, 10-meter link, over which a sender can transmit at a rate of 150 bits/sec in both directions. Suppose that packets contain- ing data are 100,000 bits long, and packets containing only control (e.g., ACK or handshaking) are 200 bits long. Assume that N parallel connections each get 1/N of the link band- width. Now consider the HTTP protocol, and suppose that each downloaded object is 100 Kbits long, and that the initial downloaded object con- tains 10 referenced objects from the same sender. [2 marks] [2 marks] [3 marks] (e) i. The equation for determining the download time for a peer-to- peer file sharing is given below. NF Dp2p ≥ max(F/us, F/dmin, u + ��N u ) s i=1i Explain what each of the following terms represents (for example: “The time for one peer to download a single copy of the file”) 1. F/us 2. F/dmin 3. us + ��Ni=1 ui Firstly, consider parallel downloads via parallel instances of non-persistent HTTP. Secondly, consider persistent HTTP. Do you expect significant performance gains over the non-persistent case? Justify and explain your answer (please clearly show your reasoning and summary calcu- lations). [12 marks] ii. Using the same variables (F, u, d) give an equation for the mini- mum download time for a client/server file download. [3 marks] [3 marks] Please go on to the next page. . . COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 3 of 9 L1 Gateway router Internet Cache Figure 1: Network with HTTP cache (f) Consider the network shown in figure 1. The hosts in the left network are requesting resources using HTTP from the origin servers on the Internet. 30% of web requests result in a cache hit (the request has been previously downloaded and can be served from the cache). Each host (not including the cache) generates a request every 20 seconds and each request requires the download of 1Mbit. What download bandwidth is needed on the access link, L1, in order for this HTTP download traffic to utilise a no more than 80% of the link download capacity? Please go on to the next page. . . [Total for Question 1: 29 marks] [4 marks] COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 4 of 9 Transport Layer Question 2 (a) Answer True (T) or False (F). Please note that each wrong answer will incur a penalty of 1 mark so guess at your own risk. Please only write the number of the statement and T or F in your answer books. 1. Each TCP socket is identified with a 2-tuple: source port number; and destination port number. 2. Host A is sending Host B a large file over a TCP connection. As- sume Host B has no data to send Host A. Host B will not send acknowledgements to Host A because Host B cannot piggyback the acknowledgements on data. 3. Suppose Host A is sending Host B a large file over a TCP connec- tion. The number of unacknowledged bytes that A sends cannot exceed the size of the receive buffer. 4. Suppose Host A sends one segment with sequence number 38 and 4 bytes of data over a TCP connection to Host B. In this same segment the acknowledgement number must be 42. 5. With the SR protocol, the receiver will only acknowledge packets within it’s current window. 6. With GBN, it is possible for the sender to receive an ACK for a packet that falls outside of its current window. (b) Why does TCP include a deviation in its timeout calculation? Why doesn’t it just use the Estimated RTT? (c) Consider a reliable data transfer protocol that uses only negative ac- knowledgements (NAK). i. Suppose the sender sends data only infrequently. Would a NAK- only protocol be preferable to a ACK-only protocol? Why (think about the receiver’s view and recovery from a failed transmis- sion)? ii. Now suppose the sender has a lot of data to send and the end-to- end connection experiences few losses. In this second case, would a NAK-only protocol be preferable to a protocol that uses ACKs? Why ? iii. Given a choice of the ACK-only scheme in TCP and the NAK-only scheme just described which scheme will you use if you protocol needed to work under high data rate and low BER conditions? [12 marks] [4 marks] [3 marks] [2 marks] [2 marks] Please go on to the next page. . . COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 5 of 9 (d) Slow-start and Congestion avoidance are two phases of TCP’s conges- tion avoidance mechanism. i. At what point does TCP change from slow-start to congestion avoidance and how is this value determined? ii. Why should three duplicate ACKs trigger fast retransmit/fast re- covery instead of dropping back to slow start? Please go on to the next page. . . [3 marks] [2 marks] [Total for Question 2: 28 marks] COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 6 of 9 Routing and Internet Protocol Question 3 LAN adapter FBGI AH 1 543 6 2 E C (router) DJ Figure 2: LAN network diagram (a) You are asked to assign IP addresses to all of the interfaces found in the network identified by A, B, C, D, E, F shown in Figure 2. In your assignment you must make sure that you do not waste any IP addresses given that you only have five hosts and a router. Please clearly show your work to justify the IP address assignment and write down three IP addresses for A, B, and C. (b) Consider a datagram network using 8-bit host addresses. Suppose a [4 marks] router uses longest prefix matching (like the example in lecture where the organisation changed providers) and has the following forwarding table: Prefix Match Interface 10 10 1 111 2 otherwise 3 For each of the four interfaces, give the associated range of destination host addresses and the number of addresses in the range. Please go on to the next page. . . [6 marks] COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 7 of 9 (c) The Figure 3 shows a three-node network that uses a distance-vector protocol for routing. B 1 A7C 2 Figure 3: A three-node network showing bi-directional link costs i. Assuming that 999 represents the value ‘infinity’, what are the initial routing tables for each node? ii. Using the distance-vector protocol, calculate and show the final, converged, routing tables for all nodes. You should clearly show all working – marks will be deducted if you fail to show your working. (d) What network characteristic(s) would make distance vector a better choice than link state route determination? (e) When IPv4 fragments a packet, how does it know which fragments belong to the original packet when it has to reassemble them? Please go on to the next page. . . [3 marks] [12 marks] [2 marks] [2 marks] [Total for Question 3: 29 marks] COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 8 of 9 Switching and Link Layer Protocols Question 4 (a) Consider the LAN in Figure 2 i. How many subnets can you find? Write down the labels (letters) of the hosts in each subnet. ii. What is the physical topology of the subnet(s)? iii. How does the switch know A is reachable via interface 1, B reach- able via interface 2 and so on (Use a simple example to briefly explain your answer)? (b) Explain how a packet enters a Multi-Protocol Label Switching (MPLS) network, how it is routed through the network and how it is ultimately delivered. Consider using a simple diagram to help you explain your answer. (c) Suppose two nodes start to transmit at the same time a packet of length L over a broadcast channel of rate R. Now, denote the propa- gation delay between the two node as dprop. Will there be a collision if dprop < L/R? Explain why or why not. (d) Suppose the information portion of a packet contains 10 bytes consist- ing of the integers 1 through to 10 where each integer is represented using a single 8-bit unsigned binary number representation. Compute the Internet checksum for this data. (e) You have received the following data and crc code: 100110001. The generator being used is: 1101 Is there an error in this data or has it been received correctly? Show your work. Please go on to the next page. . . [2 marks] [1 mark] [3 marks] [6 marks] [2 marks] [4 marks] [5 marks] [Total for Question 4: 23 marks] COMPSCI 3001,7039 Primary Examination, Semester 1, 2016 Page 9 of 9 ICMP and Security Question 5 (a) Explain how traceroute uses ICMP in order to trace the routers along the path to a destination (b) What are the differences between message confidentiality and mes- sage integrity? (c) Suppose N people want to communicate with each of N-1 other peo- ple using symmetric key encryption. All communication between any two people, i and j, is visible to all other people in this group of N, and no other person in this group should be able to decode their com- munication. How many keys are required in the system as a whole? Now suppose that public key encryption is used. How many keys are required in this case? End of exam [5 marks] [2 marks] [4 marks] [Total for Question 5: 11 marks]