Name: Personnr: Course: [ ] 1DT131 (DV, 4p) [ ] 1DT633 (MN, 5p)

Transcription

1 Exam in omputer Networks ate ourse code T3 (V) / T633 (MN) Exam time 5 hours Problems 5 (out of 8), Total 50 points Teacher Lars-Åke Larzon Phone , Aid English dictionary Name: Personnr: ourse: [ ] T3 (V, 4p) [ ] T633 (MN, 5p) Indicate what problems you have handed in solutions to with an X in the hand-in column. learly indicate what problems you have chosen to exclude in the rightmost column. Problem Hand-in Grading [ ] Excluded from exam 2 [ ] Excluded from exam 3 [ ] Excluded from exam 4 [ ] Excluded from exam 5 [ ] Excluded from exam 6 [ ] Excluded from exam 7 [ ] Excluded from exam 8 [ ] Excluded from exam TOTAL Grade Include this page with your answers

2 Exam in omputer Networks ate ourse code T3 (V) / T633 (MN) Exam time 5 hours Problems 5 (out of 8), Total 50 points Teacher Lars-Åke Larzon Phone , Aid English dictionary This exam contains 8 problems, but no more than 5 can be handed in. hoose any 5 problems and clearly specify what problems you have chosen to exclude at the front of the envelope You must answer the questions instructively. Only one problem/sheet. If details needed to solve a problem have been left out, make reasonable assumptions and present them in your solution. You may answer in English or Swedish. Exams will be graded before August 27th. A pass grade requires 25 points, distributed over at least 4 of the chosen problems. Good Luck! Good luck!

3 . Small questions a) What does the bandwidth delay product represent? b) Explain the difference in how forwarding is done in circuit- and packet-switched networks respectively. c) Briefly outline how the size of the sending window in TP Reno is regulated. d) Give two reasons to why datagrams were chosen as the fundamental transmission unit in the Internet design. e) Why is it necessary to define a minimum size of an Ethernet frame, and what happens when the payload is below this minimum size? f) TP is a byte-oriented transport protocol. Explain how the byte orientation of TP is reflected in the service provided to the application layer. 2. Ethernet Two machines A and B are attached to a 0 Mbit/s SMA/ Ethernet with 600m cable between them. The signal propagation of the cable is m/s. At time t a = 0, A transmits the message m A to B. At time t b = µs, machine B wants to transmit the message m B to A. m A contains an IPv4 datagram with 20 bytes IP payload, while m B is a maximum-sized Ethernet frame. In case of a detected collision, A will always choose K = 0, while B will always chose the largest possible value of K. At what times will the frames be successfully received at the destination? Use 0 7 instead of to represent 0 Mbit/s to simplify the calculations. For full points, you need to present a time line showing in what order different key events occur so that it is clear how you have obtained the answer. (0p) 3. Security a) What is the purpose of a nonce and how is it used? b) escribe how a cryptographic checksum can be used for authentication. c) How can a NAT box improve security? d) What are the advantages of using a cryptographical message digest instead of public-key encryption with the private key for digital signing purposes? e) Worms are becoming more and more common in the Internet. Assuming the risk of being infected is independent of the OS you are using, give two suggestions of how to reduce the risk of being infected.

4 /24 NS /24 WWW a.org / / /24 router / /6 b.org / / / /24 HP SMTP POP HP Figure : Network setup for problem 4 4. ommunication onsider the network outlined in figure where the IP address and netmask of all interfaces are shown. The servers in the figure provides the named services. After a major power failure, all machines are rebooted and all caches used for address resolving, NS queries, routing tables etc. are cleared so that each machines don t know anything about the rest of the network other than the IP address of the default router. Now, the machine a.org wants to send some data to b.org using TP. Outline what packets/frames are sent up until the first packet containing the actual data. For each packet/frame, you don t have to specify the exact format and content. However, you must specify the source, destination and the function/intent/content of the packet/frame. (0p) 5. The network layer a) What two mechanisms help improve the distance-vector algorithm by removing routing loops that only involve two nodes? b) IP can perform two different types of source routing. What are these called, and what are the difference between them? c) Multicast addresses are chosen randomly when a multicast group is created. Given this, how can a router distinguish between multicast and unicast addresses? d) Explain how the netmask is used when doing forwarding. e) Outline how IP fragmentation and reassembly works.

5 Node A: Node B: est Nexthop ost est Nexthop ost B B 4 A A Node : Node : est Nexthop ost est Nexthop ost Node E: est F Nexthop F Node F: ost est Nexthop ost 5 E E 2 Figure 2: Network setup for problem 4 6. Routing a) The ambitious students John and Eliza have discussed how Routing can be done in the Internet and present the following suggestion to their teacher: When a client wants to communicate with the server, a full route lookup is done as usual. When the packet is sent to the server, the incoming interface in each router is recorded and saved in the packet so that it is recorded exactly what way the packet has taken through the Internet. When a packet is to be sent back from the server to the client, no route lookups but instead the last recorded path is provided with the packet. Each router can now check the list of interfaces to find out which interface to forward the packet on. At the same time, the incoming interfaces of the reverse path is recorded in the same way as before. When the first packet from the server to the client arrives, you have two interface lists representing the path from the client to the server and back again. These lists are used throughout the rest of the session to eliminate the need for costly route lookups. Ignoring the cost of recording and reading the interface lists at each router and the extra space needed for it, explain why this is still not a good idea. b) All nodes in a network with 6 nodes labelled A F and bidirectional links between the nodes have been down, and were recently been rebooted. The current contents of the routing tables are shown in figure 2. Show the contents of the routing table at after each of the following events:. E sends its routing table to its neighbors 2. B sends its routing table to its neighbors 3. B sends its routing table to its neighbors 4. F sends its routing table to its neighbors (4p) c) Suggest an order in which the nodes E, A, B and F notify their neighbors so that the entries of the routing table at reaches their correct value as soon as possible with a minimum of route flapping. d) What is the difference between routing and forwarding? Flapping in this context is when a route to a node changes due to a notification of a shorter route.

6 7. Reliability There are several different approaches to achieve reliable communication in different levels of the network. Reliability in this context means reliable transfer of data, including retransmissions when needed. a) Outline how reliability is implemented in TP. How are lost segments detected, and when can a segment be considered successfully received? b) Outline how reliability is implemented in Ethernet. How are unsuccessful transmissions detected and how are retransmissions handled? When can a frame be considered successfully received? c) Outline how reliability is implemented in Wireless 802. LAN:s. How are unsuccessful transmissions detected, and how are collisions avoided? d) Outline how reliable flooding is implemented in a link-state routing protocol. e) IP is the central component in the Internet protocol stack, still it lacks reliability. Explain why it is designed like this. 8. High-speed networking On April 4, SUNET broke the Internet land speed record by transferring over 800 Gbyte of TP data 6,000 km during 27 minutes. This corresponds to a throughput of 4.23 Gbit/s and 69,073 pentabitmeter/second. a) The RTT for the transmission was roughly 300ms. How large must the sending window of TP be to fill the pipe? b) One optimization the team did to improve performance was to increase the initial TP window size to 30,000 bytes. Explain why this optimization will increase the performance. c) Another optimization was to increase the interface queue from the default value 50 to 20,000. Explain why this optimization will increase the performance. d) The transmission traversed the normal Internet from San iego to Luleå. The hardware at the end nodes were 2.0 Ghz P:s with 52 Mbytes of RAM and 0 Gbit/s network interfaces. Where is the bottleneck and why? e) What peak througput does an average throughput of 4.23 Gbit/s correspond to, i.e., how fast could you probably transmit data if the congestion window was constant?