G. Levine 12/99 Computer Networks Sample Final to cover chapters on Logical Link Control Layers in Local and Metropolitan Area Networks Network Layer Transport Layer Possibly Application Layer/ Encryption/ Home page Issues 1) What are the functions of a connection-oriented transport layer? Choose a specific transport layer protocol and describe how it supports those functions (include fields of its header in the discussion). DISCUSSION: to provide a reliable end-to-end transport independent of the lower layers TCP Setting up and releasing a connection: SYN, ACK, FIN bits discuss the protocol discuss reliable setting up of connection with three way handshake large sequence space, acknowledgement space, time-to-live field Sequencing - sequence numbers Error control - checksum, acknowledgements, time-out at sender reset field Option Negotiation - option field (header length to support option field) Addressing - source and destination PORTS, not machines Flow control - dynamic window size QoS - option field Possible questions: TCP establishes a connection using which flags? How are sequence numbers used by TCP in establishing a connection? How do sequence numbers assist in preventing the acceptance of a duplicate packet? How does the time to live field (from IP) assist in preventing the acceptance of a duplciate packet? What are TCP addresses called? (two possible answers) 2) What are the functions of a network layer? Choose a specific network layer protocol and describe how it supports those functions (include fields of its header). ex: ATM/IPv4/IPv6 4) For the following network, use Dijstra's shortest path algorithm to find the shortest path from A to F. Include all computations at each node during the traversal. A -----4-------B -----1-----------C | \ | | 3 \ 6 1 | \ \2 \ \ | | D--------4---- E -------5---------F DISCUSSION: Mark Source and its distance as permanent [A,0], add adjacent routers (B4, D3, E2) choose the one with the shortest distance D Permanent Nodes [A0,E2] Add adjacent routers (B4, D3, F7 (total 2+5)) Permanent Nodes [A0, E2, D3] Add adjacent routers (B4, F7) no new ones adjacent to D Permanent [A0, E2, D3, B4] Add adjacent routers (C5,F7) Permanent [A0, E2, D3, B4, C5] " [A0, E2, D3, B4, C5, F6] F adjusted from path thru C shortest path from A to F is ABCF 6) Discuss an anomaly that can occur when routing through the Internet and methods for its control. DISCUSSION: We discussed message ping-ponging (ring-a-rong-a-rosie), in which packets loop during heavy traffic or downed routers, if the routing algorithm does not quickly detect the crash. Solutions include global routing algorithms, hierarchical routing, time-to-live fields, etc. We discussed duplicate packets that might be accepted for connection establishment and data transfer. A three-way handshake with a large sequence (and ack) space and a time-to-live field will prevent a duplicate packet from being accepted (explain why), and will help control the incorrect acceptance of a connection request from a duplicate CR. Fill in the best answer 8) An advantage of a connection-oriented network service over a connectionless one is that it a) provides a more reliable transmission b) provides faster response for network control packets c) provides more flexibility in routing packets d) needs less recovery time if a switch goes down e) can handle priority traffic ANSWER : a 9) ATM provides a) only a connectionless service b) variable size packets called cells c) fixed size packets called cells d) large size packets called cells e) all of its data switching on the data link layer (e is correct for frame relay) ANSWER: c 10) Transparent bridges route their packets using a) the Bellman-Ford shortest path algorithm b) distance vectors stored at each bridge c) paths determined by the source nodes and included in each frame d) a routing control center (such things existed in networks such as Tymnet) e) a spanning tree algorithm ANSWER: e 11) An advantage of bridges over repeaters is that they a) operate at the physical layer b) are faster, since they do less processing than repeaters c) can connect different types of LANs together d) can recover from most transmission noise e) can amplify signals ANSWER: c 12) A disadvantage of a hub implementing a physical STAR is that it a) generally requires more wiring than a bus or a ring b) is less reliable than a bus or a ring c) can only be used with twisted pair d) can only handle data rates of 10Mbps and below e) cannot have any intelligence built into it ANSWER: a 13) IEEE 802.3 (CSMA/CD) provides a) controlled access to the channel through tokens b) random access to the channel when the channel is idle c) controlled access to the channel through reservations ANSWER: b 18) Connections are guaranteed to be correctly released by providing in the protocol for a) A single disconnect packet b) The acknowledgement of the disconnect c) A three-way handshake d) None of the above ANSWER: d 19) IP4 addressing provides for a) Two classes of addresses b) Fixed size address fields of 32 bits c) Dynamic size address fields d) Port addresses e) Fixed size address fields of 48 bits ANSWER: b