Computer Science 825 Quiz 1 Name _______________________ Terminology: the very highest layer is the APPLICATION LAYER = Layer N the very lowest layer is the PHYSICAL LAYER = Layer 1 1. a. In the ISO/OSI model at which of the seven layers is global interoperability required. b. Which is the lowest layer that is present in hosts, but is not present in routers 2. Protocol headers used by a layer N are handled by layer K (K < N) as follows: a. The are removed before b. They are treated as user data. the packet arrives at Layer K c. They are deleted by layer K. 3. Two things that a protocol can do to earn the term "unreliable" are (1) deliver packets with bits that have been inverted and (2) deliver packets out of order. Identify two others: a. b. 4. Suppose a PCM scheme supports a bit rate of 80,000 bps and quantizes the signal into 256 distinct levels. What is the highest frequency that it can carry? 5. What bandwidth is required to support a bit rate of 100,000 BPS on a channel with a 40dB SNR? 6. How many distinct signalling states must be used to send 40,000 bits/sec over channel with a bandwidth of 4,000 HZ. (Hint: fractional #'s of signalling states don't work well!) Recall the Bell 212A modem discussed in class. It signalled using phase modulation with 4 phase angles, 45, 135, 225, and 315 degrees. 7. When a long burst of 00000000000000000000 is transmitted, a. The modem will emit a b. There will be a +45 degree pure sine wave of the form phase shift in the signal f(t) = sin(wt + 45dg) every bit time. for the duration of the burst c. There will be a +45 degree phase shift every other bit time. 8. For a signaling scheme to be properly classified as self-clocking it must: a. use phase modulation b. change state every bit time c. change state every baud time d. modulate at least two of amplitude, frequency, and phase. 9. Suppose a PCM transmission system has a bit rate of 2^20 bps and can carry frequencies between 0 and 2^12 Hz. Into how many discrete levels can the input signal be quantized by the A-D converter. 10. The main reason digital channels have much lower error rates than analog when used over long distances is: a. Digital signalling is baseband c. Digital signals are amplified instead of analog. rather than regenerated. b. Digital signals are regenerated d. Digital channels have much higher rather than amplified. SNR's than analog. 11. Do protocols (as the term is normally used in networking, e.g in SDLC, TCP, or IP) define the rules under which a layer communicates a. With the layer above it b. With a peer layer in another node. c. With the layer below it. d. All of the above. Computer Science 825 Quiz 2 Name _______________________ 1. The use of very small slot sizes in a TDM multiplexing scheme is: a. Inefficient and shouldn't b. Generally shouldn't be used for be used in either STDM or ATDM SDTM but is OK for ATDM c. Generally shouldn't be used d. Perfectly OK for both for ATDM but is OK for STDM 2. For a slotted aloha system for each value of "S" (except the S corresponding to G = 1.0) there are two different values of G which will result in the throughput being "S". For a given "S" the average delay before a packet is successfully transmitted is: a. The same for both "G"s b. Larger for the small "G" than for the big "G". c. Larger for the big "G" d. Totally independent of than for the small "G". the "G" value. 3. If packets are generated at random (exponentially distributed) by each station on a shared channel, then statically assigning 1/Nth of the bandwidth to each station (rather than having a single waiting queue) a. Will have no appreciable effect b. Will cause the delay to increase on the queueing delay. by a factor of N c. Will cause the delay to be reduced d. Will cause the delay to increase by a factor of N by N seconds. 4. Suppose a network has links that carry data at a rate of 10,000 bits per second. Suppose Node A and Node B are connected by a path that requires 6 hops (transmissions). a. What is the minimum time required to send a 400,000 bit message from A to B if message switching is used. b. Suppose packet switching is used and no additional bits are required for headers. How long will it take to send the entire message if 4000 bit packets are used. 5. Suppose headers of length H bits are used in the packet switching scheme described above. Suppose packet size is P (where P now includes header+data bits). Under this new scheme the time to send a complete packet is: a. 10000 / (P + H) b. 10000 / (P - H) c. 10000 * P d. P / 10000 6. Under the new scheme the total number of packets that must now be transmitted is: a. 200,000 * P b. 200,000 / P c. 200,000 / (P + H) d. 200,000 / (P - H) 7. Suppose the value "t" represents the ONE WAY signal propogation delay between two stations on an ethernet and that the two stations transmit and cause a collision. Suppose station 1 began transmitting first and station 2 began transmitting second. Hint: the correct answers can be drawn (w/ replacement} from the following "urn": {0 t/2 2t 8t} \ t/4 4t / \____t____/ a. What is the shortest possible amount of time after station 1 starts transmitting that station 1 can hear the collision? b. What is the longest possible amount of time after station 2 starts transmitting that station 2 can hear the collision? c. What is the shortest possible amount of time after station 2 starts transmitting that station 1 can hear the collision. 7. In a dual persistent protocol a station transmits with prob p when it initially sense a channel idle and with prob q when it has to wait for a busy channel to become idle. The proper relationship between p and q is: a. p > q b. p < q c. p = q d. q = 1 and q < p 8. Measurements of an infinite user slotted aloha channel show that in 60% of the slots there is either a successful transmission or there is a collision. a. What is the total load (mean number of transmission attempts per packet time) = G. b. What is the throughput (S). c. Is the current load below or above the load required to obtain maximum throughput. Computer Science 825 Quiz 3 Name _______________________ 1. The main reason that an ethernet channel can support more packets / packet time throughput than a satellite based aloha system is that: a. The signal travels faster thru b. All stations hear every ethernet an ethernet cable than thru transmission but only the dest space. station hears an aloha transmission c. The signal propogation delay d. Only one station at a time is is far shorter in an ethernet allowed to transmit in an aloha than in a satellite net. network. 2. The main function of the CD portion of a CSMA-CD protocol is to: a. Reduce the probability that b. Reduce amount of wasted time a collision will occur. when a collision occurs. c. Ensure that no collisions d. Both a and b. occur at all. 3. The objective of the CS part of the CSMA-CD protocol is to: a. Reduce the amount of time b. Reduce the number of caused by collisions that collisions that occur do occur c. Reduce the amount of idle d. All of the above. time in which all stations are silent 4. If K stations share a CSMA bus and the probability that any station will transmit on a given contention slot is "p", then the probability that a successful transmission will take place in a slot is: a. p b. k * p * (1 - p) ^ (k - 1) c. k * p * (1 - k) ^ (p - 1) d. None of the above. 5. The probability of a successful Tx occuring during a slot is maximized if p = a. k b. 1/e c. 1/k d. None of the above 6. If the probability computed above is denoted A, then the probability that a contention interval will last exactly j slots is: a. A ^ (j - 1) * (1 - A) ^ j b. A / j c. j * (1 - A) ^ (j - 1) d. None of the above. 7. Suppose that the average length of a contention interval on a CSMA bus is C and that the average time to transmit a packet is P. The efficiency of the channel is: a. P / C b. (P + C) / P c. C / (P + C) d. None of the above 8. Each time an IEEE 802.3 Ethernet station suffers a collision (without having a successful transmission) the mean delay until its next attempt a. Increase by tau (the one b. Increases by 2 * tau way prop delay). c. Doubles d. Increases by a factor of 2 ^ n where n is the number of collisions 9. Answer the following T or F. (Assume only the parameter indicated is being varied and all of the others remain fixed in each case. This question refers to a "traditional" shared bus ethernet.) ___ a. Increasing the average packet size will tend to make the efficiency of a CSMA CD LAN higher. ___ b. Increasing the physical length of a CSMA CD LAN will increase the length of the contention interval. ___ c. Decreasing the bit transmission rate from 10 Mbps to 1 Mbps increase the efficiency of a CSMA LAN 10. For the IEEE token bus match the control packets with the conditions underwhich they are used. (Some answers may or may not be used multiple times and some may or may not be used at all.) ____ a. A station recognizes a WHO_FOLLOWS 1. SOLICIT_SUCCESSOR_2 packet specifying its predecessors address. 2. SET_SUCCESSOR_2 ____ b. A station wishes to leave the net. 3. WHO_FOLLOWS ____ c. A station transmits the token twice 4. RESOLVE_CONTENTION without hearing its successor transmit a token or a packet. 5. CLAIM_TOKEN ____ d. A station wishes to invite a new 6. SOLICIT_SUCCESSOR station to enter the network. 7. SET_SUCCESSOR 8. WHO_KNOWS 11. Starvation of all priority classes except class 6 can occur on an IEEE 802.4 token bus: (N = # of stations). a. Never b. Any time all stations have a large backlog of class 6 traffic c. Only if THT >= TRT 4 d. Only if N * THT > TRT4 12. When a token ring interface is operating in receive mode bits flow: a. into one of the onboard b. through the store and forward receive buffers buffer and on to the next station c. into the "bit-bucket" d. Both a. and b. Computer Science 825 Quiz 4 Name _______________________ 1. In a token ring on which of the following should the "Monitor" bit NEVER be set. (Check all that apply) ___ a. A high priority data frame. ___ b. A high priority token ___ c. A 0 priority token. 2. What is the bandwidth in Hz of the range of frequencies that corresponds to all wavelengths between 1.2 x 10^(-6) m and 1.4 x 10 ^ (-6) m. 3. Using a diagram, describe the "exposed station problem" that occurs in wireless LAN's 4. Suppose a 4 chip code is used in a CDMA system. Give a single codeword that is orthogonal to -1 +1 +1 -1 5. Simultaneous transmissions of data can occur simultaneously and non-destructively on which of the following LANs (Check ALL that apply) a. An ethernet using "shared b. A switched ethernet. bus. c. A token bus d. A token ring e. An ATM LAN 6. In the hybrid MAC approaches used in satellite networks the typical objective is to act like: a. STDM at low loads and b. Aloha at low loads Aloha at high loads and STDM at high loads c. Aloha at very high and low loads and STDM at medium loads 7. In an 802.5 token ring system suppose the following events occur: Station D issues a priority 2 token Station A transmits to E using priority 2 token Station C makes a priority 3 reservation Station A generates a priority 3 token Station B transmits a priority 5 message to station C Station B transmits a priority 5 token Station C transmits a priority 5 message to station E Station C issues a priority 5 token. (and there is no more (high priority traffic that anyone has to send.) a. What is the priority of the next token issued (after (C,5))? b. Who issues it? 8. Why is it that the RTS/CTS protocol IS a good thing to do in 802.11 wireless LANS but is not useful or good in 802.3 Ethernet LANs? 9. The RTS/CTS protocol is of primary benefit with: a. DCF b. PCF c. Equal benefit in both. 10. In what order do the following interframe slots appear in an 802.11 LAN. ___ a. DIFS ___ b. SIFS ___ c. PIFS 11. The primary motivation for the use of fragmentation in 802.11 but not in 802.3 (ethernet) is: a. Ethernet has a higher b. 802.11 uses much larger bit rate packets c. 802.11 has a much higher bit error probability Computer Science 825 Quiz 5 Name _______________________ 1. In the standard bit stuffing the loss of an end flag (framing character) can be detected by: a. The appearance of some 0's b. The appearance of some 1's where where there should only be 1's there should only be 0's. c. The appearance of all 1's d. The appearance of all 0's where where there should be a mix of there should be a mix of 1's 1's and 0's and 0's. 2. Suppose a 7 bit codeword with ODD parity is being used in a single bit hamming ECC. Suppose the following codeword is received. 1 1 0 0 0 1 1 -------------------- 1 2 3 4 5 6 7 Which (if any) bit is in error? 3. a. Suppose a single bit Hamming ECC is to be performed on a 511 bit codeword. How many bits are in each of the parity checked groups (include the parity bit itself). b. How many bits does any pair of parity checked groups have in common with each other? c. How many distinct parity groups does bit 62 belong to. d. Suppose a two bit error occurs. Circle the correct description of the result. 1. The fact that there is a two error can be detected, but the error can't be corrected. 2. The two bit error will be undetected and the value that results from the correction process will be a legal code word but it will be the wrong one. 3. The two bit error will be undetected and the corrected value will be either the wrong codeword or it might be the correct codeword. 4. The two bit error will be undetected and the corrected value MAY be an illegal code word. 4. Suppose 2000 bit code words are to be transmitted and it is desired to use a Hamming code to correct error bursts of up to 100 bits. How many of the 2000 bits must be used for ECC. (Hint: Fractional bit based strategies are not likely to be rewarded!) 5. Suppose a generator polynomial of G(x) = x^4 is used. G(x) can detect a. Some, but not all, of the errors b. All errors involving the involving the last 4 bits of the last 4 bits of the message message some but not all of the but no errors that do not errors that don't involve the last involve the last 4 bits. 4 bits. c. All errors involving any of d. All single bit errors but no more the last 4 bits of the message than that. and some errors not involving any of the last 4 bits. 6. Suppose a 12 bit code word is used. How many code words are at a Hamming distance of exactly 2 from a given code word. a. 12 b. 144 c. 132 d. none of the above 7. Suppose a codeword contains "n" total bits. It is broken down as "m" msg bits and "r" ECC bits. Which of the following correctly characterizes the relationship between n, m, and r for single bit error correction to be done: r m n m n a. 2 * 2 <= 2 b. (m + 1) * 2 < 2 r m c. (n + 1) * 2 < 2 d. None of the above. 8. Suppose a BISYNCH message consists of the following byte pattern while it is "on the wire". Write down the data part of the message AFTER the receiver removes the "stuffed" bytes. | this is the data part of the msg | ------------------------------------------------------------------ | DLE | STX | 'A' | DLE | DLE | 'B' | DLE | DLE | ETX | DLE | ETX | ------------------------------------------------------------------ 9. Suppose A generator polynomial of x + 1 is used. Answer the following T or F ___ a. Some but not all double bit errors in which the 2 error bits are adjacent in the message will be detected. ___ b. All errors consisting of 5 bits in error can be detected. ___ c. All double bit errors in which the 2 error bits are not adjacent in the message will be detected. Computer Science 825 Quiz 6 Name _______________________ 1. For the shift register shown above: a. What generator polynomial is represented b. What will be the NEW contents of the shift register after an input bit of "1" is processed. c. Does the shift register represent the standard long division or pre-multiplying variety? 2. Suppose a premultiplying shift register is initialized with the first 4 bits of the message instead of all 0's. Which statement best characterizes the situation a. It works fine either way b. You will always get the wrong answer if you do it this way c. Whether or not you get the wrong answer depends on the number of 0 bits at the start of the message 3. In protocol 2, ACK's were used but no sequence numbers were used on frames or ACK's. Which of the following problems could occur. a. Lost frames would go b. Lost ACK's would result in undetected. lost frames c. Lost acks would result in d. Lost frames would result in duplicate frames. duplicate frames. 4. In protocol 3, frames's had sequence numbers but ACKS did not. Protocol 3 could fail in the presence of: a. ACK's that were lost b. ACK's that were excessively delayed c. Frames that were lost. d. Either a or b 5. When protocol 3 failed, the failure resulted in: a. The loss of 2 frames b. The loss of 1 frame and the duplication of 1 frame. c. The duplication of 2 d. The loss of 1 frame frames. 6. Consider the following simplex connection.. Since the connection is simplex the sender side needs only nxtack and nxtsend state variables and the receiver side needs only nxtrecv. Assume that: the "ACK next frame expected" convention is used; the protocol uses sender and receiver window size = N; sequence numbers never wrap; and NO timeout and retransmits occur. nxtack --------------------> nxtrecv nxtsend Answer the following T or F. At any instant in time: ____ a. nxtack must always be <= nxtrecv ____ b. nxtrecv may be < nxtack ____ c. nxtsend may be < nxtrcv 7. Suppose in a connection such as described in question 6 that after the connection has been idle for a while with nxtsend=nxtack=nxrecv=0 and the sender sends a burst of 5 packets. Which of the following values will reach "5" last a. nxtsend b. nxtrecv c. nxtack d. They will all reach 5 at the same instant 8. For each of the following combinations of propogation delay and error rate identify the most suitable variation of sliding window protocol. (The most suitable choice is the SIMPLEST one that doesn't incur a big performance penalty.) ___ a. No errors but packet time 1. Stop and Wait << propagation delay ___ b. No errors but packet time 2. Receiver buffering w/ out of >> propagation delay. order retransmission ___ c. High errors and packet time << propagation delay 3. Receiver buffering w/ go back N sender 4. Go back N. 9. For a sliding window protocol to function correctly it is necessary and sufficient that the timeout interval be greater than: a. 0 b. the one-way propagation delay c. the two-way propagation d. two times the two-way delay propagation delay. Computer Science 825 Quiz 7 Name _______________________ 1. Suppose that a satellite link has a ROUND TRIP propagation delay of 500 ms. Suppose bit transmission rate is 100,000 bits per second and packets are 1000 bits long. a. What will the efficiency of this channel be if stop and wait protocol is used. b. How large (in packets) must the SWS be in order to obtain 100% efficiency if the channel is guaranteed error free. (i.e. give me the MINIMUM SWS I can use). c. What is the MINIMUM size of the seq/ack number in bits that is required to support receiver buffering with out of order retransmission. 2. Show how a sliding window protocol that uses a 2 bit sequence number, a sender window of 3 and a receiver window of 2 can fail. 3. A buffer in the protocol discussed in class cycled though a number of queues.. Associate each transition with the process that caused it. (Some answers may be used more than once or not at all). ____ 1. Read pending to network a. Network layer process input. ____ 2. Network input to link b. Queuemsg process station output c. Timeout process ____ 3. Link output to sender window. d. Writelink process ___ 4. Sender window to free e. Readlink process list 4. The sender window logic used three indexes: nxtsend; nxtqueue; and nxtack. We said these values always remained in a specific relative order. The order was a. send <= queue <= ack b. queue <= ack <= send c. ack <= queue <= send d. send <= ack <= queue e. None of the above 5. The value of the counting semaphore wpcount (that counted the number of packets waiting to be written is given by: (assume no sequence # wrap). a. queue - ack b. send - queue c. send - ack d. queue - send 6. Suppose the sequence number has 4 bits. Then the value of the semaphore numslots is: (assume modular arithmetic is properly handled) a. 16 - (queue - ack) b. 16 - (queue - send) c. 15 - (queue - send) d. 15 - (queue - ack) 7. What were the two "fatal flaws" in the sample router protocol? (This is referring to MISSING CAPABILITIES -- not broken implementations). a - b - 8. Which of the following messages has suitable semantics for the unnumbered protocol. Hint: This question is NOT asking what specific frame types are in SDLC. It IS asking what limitations are on the semantics of messages in protocols in which frames and acks carry no numbers. a. Increment current seq # b. Set current seq # to 0. c. Both a and b. d. Neither a nor b. 9. Suppose unnumbered frames were given sequence numbers and ACK numbers just like data frames and that the receiver of an out of sequence unnumbered frame would discard it just as it would an out of sequence data frame. Would this make the functions performed with unnum frames work: a. Much better and more b. Work worse and less reliably. reliably. c. Make no appreciable difference. 10. The "failure" that the unnumbered protocol is susceptible to is: a. Undetected loss b. Undetected duplication of of a frame. a frame. c. Reordering of frames d. Both a. and b. Computer Science 825 Quiz 8 Name _______________________ 1. In the SSCOP, when <> packets have been lost, a STAT packet will contain three 4-byte fields. Identify what they are in a way the indicates you understand their usage (i.e. abstract notation such as N(X) doesn't get the job done). a - b - c - 2. In the SSCOP, when the last packet to have been received successfully and in order is 57 and the the packets that have been received since then include 60 and 63 The list part of the next "stat" packet should include: a. 58 60 61 63 b. 58 61 c. 58 60 63 d. 58 61 62 3. A router in a typical "real world" network knows a. the entire path to the b. the entire path from the destination of each packet source of each packet it it processes processes c. both a. and b. d. only the next "hop" to which the packet should be forwarded. 4. The use of a RELIABLE network layer is more desirable a. with very unreliable links b. with very reliable links than with very reliable links than with very unreliable links c. has to relation to the reliability of the links. 5. The basic backward learning routing algorithm learns only for the better. A mechanism that can be used to discover changes for the worse in the topology is to: a. Use duplexed tables and b. Periodically use flooding. periodically initialize them to all dests unknown c. Add a time stamp to the d. All of the above. packets. Suppose two stations are connected by a path that is N hops long and the probability of a successful transmission over each hop is p. 6. The expected number of total transmissions to get a packet from one end to the other if a reliable network layer is used is: (The reliable network repeats transmissions at each hop until it succeeds). a. N / p b. p / N c. 1 / p d. 1 / Np 7. The expected number of total transmissions to get a packet from one end to the other if an unreliable network layer is used is approximately: (The unreliable network layer just discards failed transmissions and depends on the transport layer to start over from the beginning if a packet doesn't succeed in making the whole trip) a. 1 / p^N b. N / (2 p ^ N) c. N ^ p / 2 ^ N d. N ^ p / p 8. Associate each of the following network layers with the class of service that it provides: ____ a. ATM network layer 1. Unreliable connectionless 2. Reliable connectionless ____ b. SNA network layer 3. Unreliable connection oriented ____ c. IP 4. Reliable connection oriented 9. Of the 4 possible network service classes identified above which two are in the most widespread use today: a. 1 and 2 b. 3 and 4 b. 1 and 3 d. 2 and 3 10. Which of the following is and advantage of the use of connection oriented network service over connectionless network service: a. more flexibility in routing b. ease of providing QoS guarantees c. less disruptive recovery d. less state information to from failed routers maintain in routers Computer Science 825 Quiz 9 Name _______________________ 1. Suppose the Bellman-Ford routing algorithm is being used and station D receives the following routing updates from neighbors A, C and E Suppose DA delay = 5, DC delay = 2 and DE delay = 3. Complete the routing table for D., A C E Cost VIA A 0 2 6 | ------------ ---------- B 6 11 9 | ------------ ---------- C 2 0 8 | ------------ ---------- D 8 5 9 | ------------ ---------- E 6 8 0 | ------------ ---------- F 4 9 10 | ------------ ---------- 2. Let N = number of nodes in a network and M = the average number of links / node. a. The minimum number of routing packets transmissions at each complete SPF routing exchange within the entire network is of order: 1. M 2. N 3. N ^ 2 4. M ^ 2 5. N * M b. The minimum number of routing packets transmissions at each complete Bellman-Ford routing exchange within the entire network is of order: 1. M 2. N 3. N ^ 2 4. M ^ 2 5. N * M c. The size of a routing info packet in Bellman-Ford is of order 1. M 2. N 3. N ^ 2 4. N * M d. The size of a routing info packet in SPF is of order 1. M 2. N 3. N ^ 2 4. N * M 3. Each routing update packet propagates a. Across the entire net in SPF b. Only to adjacent neighbors in SPF and in Bellman-Ford and Bellman Ford c. Across entire net in SPF but d. Across entire net in B-F but only only to adjacent neighbors in B-F to adjacent neighbors is SPF 4. The routing algorithm used to transmit routing updates in an SPF network is typically: a. Static tables built by b. Backward learning sysadmin c. Bellman-Ford d. Flooding 5. Routing loops can occur in SPF a. only when all nodes use the same b. only when some nodes use a cost matrix different matrix than others c. in both cases a. and b. d. Never, the SPF algorithm ALWAYS guarantees optimal routes 6. In a network that uses SPF suppose station B constructs the following cost matrix A B C D E A 0 2 - 1 6 B 3 0 6 - 9 C - 6 0 7 2 D 7 - 12 0 2 E 4 9 2 4 0 Draw the routing tree from node B to all destinations. Label the ARCs of the graph in the order in which they are made permanent in the SPF algorithm. 7. When it was originally used (in the late 1970's) for routing in the internet, the Bellman Ford's metric computation and routing exchange algorithms were: a. Average queue length + b. Instantaneous queue length exchange at least once a and exchange every 1 minute minute c. Average queue length + d. Instantaneous queue length and exchange several times a exchange several times per second second. 8. When it replaced Bellman-Ford (in the late 1970's) for routing in the internet, the OSPF metric computation and routing exchange algorithms were: a. Average queue length + b. Instantaneous queue length exchange at least once a and exchange every 1 minute minute c. Average queue length + d. Instantaneous queue length and exchange several times a exchange several times per second second. 9. In the modern Internet what is the essential difference between multi-homed and transit Autonomous systems: 10. The routed and gated routing daemons are primarily use for routing a. Within an AS b. between AS's c. Used for both purposes Computer Science 825 Quiz A Name _______________________ 1. a. Suppose network addresses are 24 bits and flat routing tables are used. What is the maximum number of entries that may be required in a routing table. b. Suppose hierarchical routing is used and the addresses are partitioned as an 6 bit Area address; a 8 bit Network address and an 10 bit host addres. Now what is the maximum number of routing table entries. 2. Order the following broadcast algorithms according to the number of packet transmissions needed to complete a broadcast (1 = least transmissions; 3 = most). ______ a. Reverse Path Forwarding ______ b. Spanning tree. ______ c. Flooding 3. In the network shown suppose all hosts know the optimal route to A (hop count metric) and A initiates a reverse path forwarding broadcast. How many unnecessary transmissions will occur? 4. Suppose a wide-area multicast protocol is in effect and a router receives a message indicating that a host to which it was forwarding the multicast no longer wants to receive it. That request should be forwarded upstream toward the source of the m'cast if: a. there are no more hosts downstream b. there are additional hosts of the router who are receiving the downstream who continue mcast. to receive it. c. it must always be forwarded upstream d. it should never be forwarded upstream 5. Preallocation of resources is a practical mechanism for congestion avoidance a. Only for reliable connnection b. for either reliable or unreliable oriented network layers connection oriented networks c. Only for connection-less network d. Works fine for all of the above layers 6. The main factor that determines whether or not packet-discarding can be used as a congestion control mechanism in a network layer is: a. Whether the network layer b. Whether the network layer is is connection oriented or reliable or unreliable connection less. c. Both are equally important. d. Neither is important at all. 7. Suppose that "leaky bucket" and "token bucket" admission control algorithms support identical maximal long-term sustained admission rates. Which of the following best characterizes their response to short high-rate arrival bursts: a. Leaky bucket allows short b. Token bucket allows short term admission rate to exceed term admission rate to exceed long term rates but token bucket long term rates but leaky bucket does not. does not. c. Both allow short term admission d. Neither allows short term rates rates to exceed long term rates. to exceed long term. 8. Suppose three traffic arrival streams have the same mean arrival rate. Order the streams from 1 (smallest queue length) to 3 (largest queue length) according to the the following characteristics: ____ a. High variance ... no self-similarity ____ b. High variance ... high degree of self-similarity ____ c. Very low variance ... high degree of self-similarity. 9. "Bridging" is the interconnection of two networks at the: a. Physical layer b. Datalink layer c. Network layer d. Transport Layer 10. Aggregated on-off traffic sources have been shown to produce long range dependence as the number of sources and the length of the aggregation period becomes large if: a. the variance of the time b. the mean time between packet between individual packet transmissions is infinite transmissions is infinite c. the variance of the length of d. the effect always occurs regardless the on/off periods is infinite of the nature of the on-off processes Computer Science 825 Quiz B Name _______________________ 1. The primary objective of the spanning tree algorithm in a transparent bridge system is to: a. Reduce the number of b. Prevent indefinite packet looping. destination unknowns c. Ensure that the optimal d. All of the above. route is taken. 2. Which routing algorithm is used by the source routing bridge in routing "discovery" packets to the destination. a. Spanning tree. b. Flooding c. Backward learning. d. Shortest path first. 3. Which routing algorithm is used by the spanning tree bridge if the location of the destination node is UNKNOWN? a. Spanning tree. b. Flooding c. Backward learning. d. Shortest past first. 4. Which routing algorithm does the spanning tree bridge employ in order to minimize the occurrences of destination unknowns. a. Spanning tree. b. Flooding c. Backward learning. d. Bellman-Ford. 5. With respect to the objective of Shortest Path Routing, which of the two bridging techniques we studied is superior. a. Spanning tree b. Source routing c. Both provide optimal paths 6. Consider the bridged network shown. Circles are bridges and lines LANs. Mark each interface on each bridge "R", "D", or "N" depending up whether it is the "Root Port", "Designated bridge port for the LAN", on "NOT in the spanning tree". 7. A SNA VR manager receiving a packet with the Change Window Reply Indicator must: a. Stop sending immediately b. Set the current pacing group size to the minimum pacing group size. c. Decrement the current d. Increment the current pacing group pacing group size. size. 8. A VR manager receiving a packet with the Reset Window Indicator must: a. Stop sending immediately b. Set the current pacing group size to the minimum pacing group size. c. Decrement the current d. None of the above. pacing group size. 9. When a VR manager receives a pacing response before the pacing count reaches 0 and it doesn't contain CRWI or RWI it a. Sets the pacing count to b. Increments the current pacing the current pacing group group size by 1. size. c. Adds the current pacing d. Sets the pacing count to the group size to the pacing maximum pacing group size. count. 10. What is the function of the "protocol" byte found in the IP header? 11. Suppose the following is a diagram of an SNA network: B E / \ / \ A D G \ / \ / C F Answer the following T or F. ___ a. Nodes B and C can both have explicit routes with the identity (3, F). ___ b. For node B to have a functional explicit route with identity (3, G), node D must also have an explicit route with identity (3, G). ___ c. Nodes A and C can both have explicit routes with the identity (3, G) if A's (3,G) passes through E and C's (3, G) passes through F. Computer Science 825 Quiz C Name _______________________ 1. In IPV4 fragmentation and reassembly are performed as follows: a. Both can occur anywhere b. Source host must fragment but reassembly is done just past the "bottleneck" router c. Fragmentation may occur d. Fragmentation always occurs at anywhere but reassembly occurs the source and reassembly at only at the destination. the destination. 2. In the old class based IP address assignment architecture, which class of net does a node whose address is 190.215.17.139 belong to. a. A b. B c. C d. D 3. IP datagrams can be up to 64K in size but the fragment offset field is only 13 bits. How is it possible to know where a fragment is supposed to be placed if a datagram is larger than 8KB? 4. Which was of the following was NOT a design objective of IPV6. a. Enlarge the internet address b. Provide better support space. for mobile hosts. c. Provide additional fragmentation d. Reduce computing time and table facilities to make it more ATM size requirments in backbone compatible. routers 5. The type of service provided by IP and ATM is properly contrasted as follows: a. IP provides unreliable service b. IP provides connection but ATM is reliable. oriented service but ATM is connection less c. ATM provides unreliable service b. ATM provides connection but IP is reliable. oriented service but IP is connection less 6. The number of distinct ATM addresses (not connection id's) is: a. 1 b. 2^8 c. 2^20 d. 2^160 7. The full end-system ATM address is used in: a. SVC setup b. Routing of individual SVC cells c. PVC setup d. Routing of PVC cells 8. The number of address bytes that each ATM host "inherits" from the switch to which it is attached is: a. 6 b. 13 c. 14 d. 19 9. The connection identifier (vpi, vci) for an SVC is assigned by a. the edge switch b. the originating host O/S c. the application program 10. The connection identifier (vpi, vci) for a specific SVC a. never changes b. MUST change at each hop c. MAY change at each hop 11. In an ATM network routing decisions: a. Must be made by the sysadmin b. May be made dynamically before the network is activated. at connection setup time c. May be made dynamically on a per cell basis as in IP. 12. Which of the following is an ATM network layer not permitted to do a. Discard cells b. Deliver cells with multiple bit errors c. Reorder cells d. All of the above Computer Science 825 Quiz D Name _______________________ Suppose a 3 way handshake is being used to establish a connection. Let SYN-1 represent a packet that carries SYN and no ack. Let SYN-2 represent a packet that carries both SYN and ACK that is sent in response to a SYN. 1. A delayed duplicate SYN1 received in the ESTABLISHED state is: a. Responded to with SYN2 b. Always causes TCP to crash. c. Responded to with RESET. d. Responded to with a data packet specifying current ACK and SEQ. 2. A delayed duplicate SYN1 received in the SYN-sent state is detected because: a. Both SEQ and ACK are b. Undetected and causes the known to be wrong. recipient to respond with a SYN-2 c. The SEQ number is wrong. d. The ACK number is wrong. 3. A delayed duplicate SYN1 received in the LISTEN state is: a. Responded to with RESET. b. Responded to with a data packet specifying current ACK and SEQ. c. Responded to with SYN2 d. Always causes TCP to crash. 4. Suppose Tomlinson's clock proposal is used with a 18 bit sequence number and a clock that ticks 4096 times a second. Suppose maximum packet lifetime is 20 seconds. a. What is the MINIMUM amount of time in seconds that can elapse before sequence number resynchronization must occur. b. How often will synchronization be needed if packets are generated at 7/8 the maximum allowed rate. c. If the initial sequence number is 0 what will be the sequence number at the time resequencing takes place. (in question b). 5. Suppose a receiver in a stop and wait file transfer protocol uses an ack then write policy and the sender uses a policy of retransmitting after a crash if and only if there was an unacknowledged packet outstanding at the time of the crash. Identify whether each of the following sequence of events will result in correct behavior (OK) a lost packet (LOST) or a duplicate (DUP). AWC AC(W) C(AW) 6. Suppose a TCP connection has SND.UNA = 2600 and SND.NXT = 3600 and a packet arrives with ACK = 3200 and WINDOW = 1000. a. How many more bytes may be transmitted before the sending TCP must stop sending because it runs out of usable space in the sender window. b. Suppose the sending application has an allocation of 8000 bytes of transmit buffer space. At what sequence number will the application be "put to sleep" if no additional acks are received. 7. Suppose an application has an allocation of 8,000 bytes of TCP receive buffer space. The last byte that the application has consumed is 4000 and the present offered window is 7000. What is RCV.NXT? 8. Although the objectives of a reliable datalink protocol and an reliable transport protocol are basically the same, we identified several factors that make the implementation of the reliable transport over an unreliable connectionless network layer such as IP considerably more difficult. Identify two of them: a - b -