Sirindhorn International Institute of Technology Thammasat University

Transcription

1 Name ID Section Seat No Thammasat University Final Exam: Semester, 205 Course Title: Introduction to Data Communications Instructor: Steven Gordon Date/Time: Monday 4 December 205; 3:30 6:30 Instructions: This examination paper has 20 pages (including this page). Conditions of Examination: Closed book; No dictionary; Non-programmable calculator is allowed Students are not allowed to be out of the exam room during examination. Going to the restroom may result in score deduction. Turn off all communication devices (mobile phone etc.) and leave them at the front of the examination room. The examination paper is not allowed to be taken out of the examination room. A violation may result in score deduction. Write your name, student ID, section, and seat number clearly on the front page of the exam, and on any separate sheets (if they exist). Assume bits are ordered from left to right. For example, for the data 0000, the first (st) bit is 0 and the last (8th) bit is. Assume the speed of transmission is m/s Reference material included at the end of the exam may be used.

2 Question [22 marks] This question contains multiple choice sub-questions. You must write your answers for all sub-questions in the table below. E.g. if you think the answer for sub-question (a) is option 3 then write 3 in the table below. Select only one answer (even if you think there are multiple correct answers; selecting two or more answers automatically gets 0 marks). Correct answers are worth mark. 0 marks for an incorrect answer or no answer. Question Answer Question Answer (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) (k) (l) (m) (n) (o) (p) (q) (r) (s) (t) (u) (v) Table : Answers for multiple choice questions (a) What is the length of the address assigned to a wireless LAN device (e.g. USB adapter, or chip on the motherboard) by the device manufacturer?. 2 bits 2. 6 bits bits bits bits bits bits bits 9. Wireless LAN devices do not have addresses assigned by the manufacturer (b) Which of the following statements about internetworking is true?. A subnet has only one router attached 2. A router always has two network interfaces 3. A router cannot be a source of IP datagrams 4. A host can have only one network interface 5. A host may forward IP datagrams 6. A host may send a packet with a router as destination 7. None of the above Introduction to Data Communications 2

3 (c) Which command causes the computer to send multiple UDP packets to a computer with address...?. iperf -c... -t iperf -c... -u 3. iperf -s -w iperf -s -u 5. ping -c Both. and Both. and Both. and Both 2. and Both 2. and 5.. Both 4. and 5. (d) What is the special IP with all binary s used for?. a source address if a device is sending to itself 2. a source address when a device doesn t yet have an IP address 3. a destination address to send to all devices on the same subnet 4. a destination address to send to all devices on another subnet 5. a destination address if a device is sending to itself 6. an address of a subnet 7. an address of a router (e) To deliver s from clients to server, SMTP is used. For clients to download s from an server IMAP or POP are commonly used. Which layer of the TCP/IP layered model do these protocols belong to?. Application 2. Data Link OSI 5. Physical 6. TCP/IP 7. Transmission 8. Transport (f) With some adaptive routing protocols information about the network (e.g. topology, link costs) comes from all nodes and is updated on a periodic basis. For normal flooding, where does the information come from and how is it updated?. All nodes, periodic updates 2. All nodes, continuous updates 3. All nodes, no updates 4. No nodes, periodic updates 5. No nodes, continuous updates 6. No nodes, no updates 7. Neighbour nodes, periodic updates 8. Neighbour nodes, continuous updates Introduction to Data Communications 3

4 (g) Which layer(s) of the TCP/IP layered model do the standards for Ethernet and WiFi focus on?. Physical only 2. Data link only 3. Network only 4. Physical and data link 5. Data link and network 6. Physical, data link and network 7. None of the above (h) The process of finding a path through a network is called:. congestion control 2. connection-oriented 3. forwarding 4. fragmentation 5. routing 6. None of the above (i) Normally TCP would be implemented as:. Software in the operating system 2. An application installed by users that want to use UDP 3. Part of a user application, such as web browser or 4. Hardware on the LAN/WAN interface cards 5. Device drivers that control the LAN/WAN interface cards (j) A transport protocol is identified by:. Ethernet address 2. address 3. IP address 4. domain name 5. port number 6. protocol number (k) Which protocol provides flow control, error control and connection management:. ARP 2. DNS 3. ICMP 4. HTTP 5. TCP 6. UDP (l) The type of switching used in IP networks is:. Circuit switching 2. Datagram packet switching 3. Ethernet switching 4. Virtual circuit packet switching 5. None of the above Introduction to Data Communications 4

5 (m) Which of the following is a network layer protocol?. Ethernet 2. HTTP 3. DHCP 4. IP 5. TCP 6. UDP 7. WiFi (n) The standards for Ethernet are developed and maintained by:. IEC 2. IEEE 3. IETF 4. ISO 5. ITU 6. OSI 7. W3C (o) Which technique is described by transmitting data from multiple users across a single link but at different times?. ADSL 2. ARQ 3. CDMA 4. PCM 5. TCP 6. TDM 7. QoS (p) Select the correct set of features provided by IP:. Addressing, error control, security 2. Connection-less, addressing, data delivery 3. Addressing, flow control, quality of service 4. Addressing, error control, connection-less 5. Connection-oriented, data delivery, flow control 6. Flow control, addressing, data delivery (q) Which transport protocol allows real-time applications (e.g. video streaming) to send packets without waiting for ACKs?. ARQ 2. DHCP 3. DNS 4. HTTP 5. IP 6. NIC 7. TCP 8. UDP Introduction to Data Communications 5

6 (r) What is the shortest time it takes an Ethernet interface on a new laptop to transmit one bit?. 0. ns 2. ns ns 4. 0 ns ns ns ns ns (s) Compared to other common topologies in LANs (such as bus, mesh, ring), which of the following is a disadvantage of using a star topology for LAN?. Failure of a special node results in no stations being able to communicate 2. Poor performance due to sharing of point-to-multipoint links between users 3. Each station requires many network interfaces 4. Too many links needed in the network 5. Failure of a single station results in no stations being able to communicate 6. Failure of a single link results in no stations being able to communicate (t) If IP software on a computer creates and then sends an IP datagram with destination address , then the datagram will be:. Delivered to all devices on the same subnet as the source 2. Delivered to all devices on subnet Delivered to device with address Delivered to the source 5. Delivered to the switch 6. Delivered to the router on the same subnet as the source 7. Not delivered to any destination it will cause an error (u) A disadvantage of circuit switching (versus datagram packet switching) is:. Switches are complex because they must process every piece of data 2. It is difficult to guarantee performance for delivery of data 3. Data is likely to be dropped at switches 4. As new connections are requested they may be blocked 5. A larger delay (than datagram packet switching) will always occur when transferring the same amount of data (v) Which of the following is true?. All subnets in an internet must have the same network portion in their IP address. 2. The host portion of an IP address must be the same length across different subnets. 3. The network portion of an IP address must be the same length across different subnets. 4. All hosts in the same subnet must have the same host portion in their IP address. 5. All hosts in the same subnet must have unique network portion in their IP address 6. All hosts in the same subnet must have unique host portion in their IP address. Introduction to Data Communications 6

7 Question 2 [6 marks] Consider the network in Figure. The nodes/circles are datagram packet switches. Link 2 Link 2 5 Link 3 3 Link 4 4 Link 5 Figure : Network Topology The link characteristics are: delay to send one packet; throughput; and number of gigabytes that can be sent for Baht. The values of these three characteristics for each link are: Link : 20 ms, 300 Mb/s, 5 GB Link 2: 50 ms, 300 Mb/s, 4 GB Link 3: 0 ms, 600 Mb/s, 2 GB Link 4: 30 ms, 200 Mb/s, 3 GB Link 5: 0 ms, 400 Mb/s, 4 GB Least cost routing must be used in the network. The metric used depends on the objectives of the nodes. (a) If node wants to transfer a very large amount of data (00 s of GB) to node 5, but wants to pay as little as possible and does not care about performance, what is the path and what is the cost of that path? Explain or show how you calculated the path cost. [2 marks] Introduction to Data Communications 7

8 (b) If node wants to transfer a very large amount of data to node 5, and the cost doesn t matter, which path from to 5 is the best path to use? Explain why. [2 marks] (c) Explain how a cost can be assigned to each link so that least cost routing can be used, that would automatically select the path you chose in part (b). Your method for assigning a cost should only use the link characteristics available in this question. What would the path cost be? [2 marks] Introduction to Data Communications 8

9 Question 3 [0 marks] Consider the network in Figure 2. The nodes/circles are packet switches. E G B F K L M A C I O D H J N Figure 2: Network Topology 2 Assume flooding is to be used in the network to deliver a data packet from node D to node L. (a) What is the optimal value of the hop limit that source node D should use? Explain your answer. [3 marks] Introduction to Data Communications 9

10 Assume instead of flooding, adaptive routing is to be used in the network to deliver data packets. The routing metric is not hops. The cost of a link is identical in both directions. The following least-cost paths have already been determined: M L K G E B B E F C H I N O J H C F E B B A D G F C A D M L K F (b) Draw the optimal routing tables for nodes B and F. An optimal routing table is one with the least number of rows. Each row must contain only one value for the destination and next router, with the exception that you can use * in the destination column to indicate any value. In forwarding, assume the table is processed rowby-row: if one row matches, then the subsequent rows are ignored. [4 marks] Consider datagram packet switching being used to deliver packets on the path A B E G K. All packets must contain a 20 Byte header. They may contain a varying amount of data. The source A has 4,000 Bytes of data to send to K. (c) Explain three (3) factors that are important in node A selecting the amount of data that should be carried in each packet. [3 marks] Introduction to Data Communications 0

11 Question 4 [6 marks] Assume a 4-bit sequence number is used in a sliding window flow control protocol (that is, maximum window size is 5). The current state of a source node is: Last frame ACKed = 3 Current window size = 9 frames Then the node transmits 5 DATA frames, and then receives an ACK (Receive Ready) frame with number. After these frames have been transmitted/received, what is the new value of: (a) Last frame ACKed: (b) Last frame transmitted: (c) Current window size: Introduction to Data Communications

12 Question 5 [ marks] Consider the network topology in Figure 3. There are four IP routers, RA, RB, RC and RD. There are three hosts on the left subnet and three hosts on the right subnet. All subnets use Ethernet, with both the left subnet and right subnet containing switches (S and S2). Assume there are no hosts on any other subnets (that is, there are currently only six hosts in total). H RB 2 H4 H2 S RA 2 RD 3 S2 H5 H3 3 RC 2 2 H6 Figure 3: Network Topology 2 The numbers next to each link into a device in Figure 3 are interface numbers. For example, RA has three interfaces: interface connects to the switch, interface 2 connects to RB, and interface 3 connects to RC. Table 2 lists the MAC and IP addresses currently assigned to a selection of the interfaces. Table 2: Interface Addresses Device Interface MAC IP H :22:33:aa:bb:cc /24 H2 22:33:44:bb:cc:dd /24 H3 33:44:55:cc:dd:ee? RA 2:34:56:78:90:ab /24 RA 2 cd:ef:2:34:56: /6 RA 3 90:ab:cd:ef:2: /6 RB aa:bb:cc::22: /6 RB 2 bb:cc:dd:22:33: /20 RC cc:dd:ee:33:44: /6 RC 2 dd:ee:ff:44:55: /8 RD 56:78:90:ab:cd:ef /20 RD 2 ef:cd:ab:90:78: /8 RD 3 34:2:ef:cd:ab: /2 H4 99:a:b2:c3:d4:e /2 H5 88:f6:a7:b8:c9:d /2 H6 77:e:f2:a3:b4:c /2 The following sub-questions give a scenario of one device sending data to others. For that given scenario, you need to fill in the details of the packet (Ethernet frame and/or IP datagram), that is, the packet header field values and other information about the packet. Assume the default/typical headers are used (see the Reference Material). Introduction to Data Communications 2

13 (a) H2 has TCP data to send to H4. The entire TCP segment (including header) is 500 Bytes long. The path that the IP datagram takes is H2 RA RB RD H4. The initial time to live set by H2 is 0. Fill in the table below for the packet received by H4. [4 marks] Field Ethernet Source Value Ethernet Destination IP Source IP Destination IP Protocol IP Total Length (b) The same as for part (a) (that is, the same packet sent on path H2 RA RB RD H4), but fill in the table below for the packet sent by RB. [3 marks] Field Ethernet Source Value Ethernet Destination IP Source IP Destination IP TTL H2 wants to have data sent to all devices on the subnet that H4 is attached to. H2 has two approaches to do this: use unicast or use broadcast. (c) What is the value of the broadcast address H2 should use? [2 marks] (d) Explain two advantages of using broadcast, compared to using unicast. [2 marks] Introduction to Data Communications 3

14 Question 6 [3 marks] WiFi uses a stop-and-wait ARQ mechanism for data transfer between client (e.g. laptop) and Access Point (AP). However before each DATA frame is sent, the sender must wait w µs, and during that time, check that no other device is transmitting. For example, the client waits w µs, transmits the DATA frame, and upon receipt, the AP immediately transmits an ACK frame in reply. Then before transmitting the next DATA frame, the client waits w µs. The waiting time is to give opportunity for other devices to transmit, and it cannot be removed. A WiFi DATA frame contains a 66 Byte header and d Bytes of payload (where 00 d 500). An ACK frame is 4 Bytes of header (no payload). Apart from the waiting time w, assume no processing delays. The link propagation delay is p µs from client to AP, and p µs from AP to client. In WiFi the DATA frames and ACK frames may be transmitted at different data rates. Let r data be the data rate of DATA frames, and r ack be the data rate of ACK frames. Assume the following values: r data = 54 Mb/s, r ack = 2 Mb/s, p = µs, w = 70 µs. (a) If client has a large amount of data to send to the AP, and there are no other devices nearby transmitting (e.g. no interference) and no frame loss, then what is the maximum throughput that can be achieved? [4 marks] Introduction to Data Communications 4

15 Now consider a DATA frame with payload of d = 04. If a DATA frame is lost (or received in error at the AP), then after a timeout, the client waits w µs, and then retransmits the DATA frame. The client uses a timeout value of x µs, where the timer starts immediately after the transmission of an entire DATA Frame (i.e. after the last bit is transmitted). (b) What the smallest value of T O that the client should use? [2 marks] Assume instead of using the smallest value of T O, the client uses T O = 200 µs. Also, assume you determine that 85% of new frames transmitted will be delivered successfully to the AP, 0% of new frames transmitted will be lost and require a single retransmission, and 5% of new frames transmitted will be lost and the first retransmission will also be lost (but the second retransmission will not be lost). (c) If the client has a large amount of data to send to the AP, and there are no other devices nearby transmitting (e.g. no interference), then what is the maximum throughput that can be achieved? [4 marks] Introduction to Data Communications 5

16 Now consider a path from the client (e.g. laptop) to a web server, that goes via the AP, and then via several other links. The ARQ mechanisms on the links often ensure data is eventually successfully delivered across each link. Assume that after applying the ARQ scheme on each link in the path, every DATA frame is successfully delivered across each link. TCP is used on the client and server, and uses a Go-Back-N ARQ mechanism. (d) Explain why, even if the link ARQ mechanisms are successful, it is still important to use ARQ in TCP. [ marks] (e) Explain how Go-Back-N and Selective-Reject operate differently when after the source sends a window of frames, one of those frames is lost. [ marks] (f) What is an advantage of Go-Back-N (when compared to Selective-Reject)? [ mark] Introduction to Data Communications 6

17 Question 7 [6 marks] A sliding window protocol is used over an error free link, with a window size of 5 data frames. Each data frame has a transmission delay of 000 µs, and the header is 0% of the total size (the remaining 90% of the data frame is payload, i.e. real user data). ACK transmission time is 0 µs, while the link propagation delay is 8 ms (that is, milliseconds). Assuming the source always has data to send, what is the maximum efficiency, expressed as a percentage, that can be achieved in delivering user data across the link? Introduction to Data Communications 7

18 Question 8 [6 marks] You are testing the performance of an Ethernet link using a web browser on computer A (IP...) and a web server on computer B (IP...2). Computer A and B are connected directly via the single Ethernet link, which is configured to only use data rate of 00 Mb/s. The web server has a set of files on it for testing purposes: K.html, 2K.html, 00K.html and 00G.html. The contents of the files are not important, however the size of the file is given in the name. For example, K.html is exactly 000 Bytes in length. Every HTTP request sent by the browser is 00 Bytes of header only (no data). Every HTTP response sent by the server is 00 Bytes of header plus the requested file as data. Assume the processing, queueing and propagation delays are 0. (a) If you use the URL in the browser on computer A, what is the round trip time (also called application response time)? [3 marks] (b) If you use the URL in the browser on computer A, what is the application throughput for the download from server to browser? You may ignore any overhead due to TCP error or flow control (e.g. ACKs). [3 marks] Introduction to Data Communications 8

19 Reference Material Selected well-known ports: FTP 20 and 2 SSH 22 Telnet 23 SMTP 25 DNS 53 HTTP 80 HTTPS Bytes Version HLength DiffServ ECN Total Length Identification Flags Fragment Offset Time To Live Protocol Header Checksum Source IP Address Destination IP Address Options + Padding (optional) Data Figure 4: IP Datagram Format. Flags: Reserved, Don t Fragment, More Fragments 20 Bytes Source Port Destination Port Sequence Number Acknowledgement Number HLength Reserved Flags Advertised Window Checksum Urgent Pointer Options + Padding (optional) Data Figure 5: TCP Segment Format. Flags: CWR, ECE, URG, ACK, PSH, RST, SYN, FIN Introduction to Data Communications 9

20 Bytes Source Port Total Length Destination Port Checksum Data Figure 6: UDP Datagram Format 46 to 6 Bytes 6 Bytes 2 Bytes 500 Bytes 4 Bytes Destination Address Source Address Ether Type Data CRC Checksum Figure 7: Ethernet Frame Format Selected Protocol numbers: ICMP 6 TCP 7 UDP Introduction to Data Communications 20