1 of 10 CS 455: Computer Communications and Networking (Spring 2021) Midterm Exam Name: ___________________________________ G#: ______________________________________ Exam policy: - This exam is open book and open notes. - The exam has to be taken individually, not with a partner. Do not share your answers with others or on Piazza. - The total points for this exam are 100. 2 of 10 True or False [30pt] 1. Internet eXchange Points (IXP) are only used by tier-1 ISPs to peer with each other. True False 2. A fiber cable with higher transmission rate always has a smaller propagation delay compared to an ethernet cable with a lower transmission rate. True False 3. The default HTTP behavior is to use persistent connections and also pipelining. True False 4. HTTP GET method can be used to upload form input from client to the server. True False 5. With non-persistent connections between browser and origin server, it is possible for a single TCP segment to carry two distinct HTTP request messages. True False 6. HTTP response messages can have an empty message body. True False 7. Root DNS servers have all the mappings for all of the hosts on the Internet. True False 8. With DASH (Dynamic Adaptive Streaming over HTTP), different versions of a video are stored on the HTTP server, all of which can be accessed using the same URL. True False 3 of 10 9. Some CDN companies host their server clusters in IXPs instead of deploying them inside ISPs. True False 10. A UDP socket is uniquely identified by (source port, source IP, destination port, destination IP). True False 11. Suppose that the last SampleRTT in a TCP connection is equal to 1 second. The current value of TimeoutInterval for the connection will be necessarily >= 1 second. True False 12. Suppose Host A sends one segment with sequence number 25 and 5 bytes of data over a TCP connection to Host B. In this same segment, the acknowledgement can be 30. True False 13. A TCP receiver which receives an out-of-order segment with a higher than expected sequence number immediately sends back a duplicate ACK indicating the sequence number of the last segment correctly received in order. True False 14. Go-back-N, selective repeat and TCP, all three use cumulative ACKs. True False 15. With TCP fast retransmit, a sender can detect a (possibly) lost segment even before the time-out period. True False 4 of 10 Answer the following questions [70pt] 1. [10pt] Consider Host A is pinging Host B in the network shown below. The ping response outputs the total round trip time from A to B. The round trip time found by the ping command is 58ms. For simplicity, let’s assume that queueing delay and processing delay at all nodes are 0. The network is shown below with link transmission rates shown for each link. A--------------------R1-----------------------R2--------------------B 100kbps 25kbps 50kbps Assume that propagation speed on all links is 2.5x10^8 m/s and the length of the ping packets (request and response) is 25 bytes. The length/distance of all three links (A-R1, R1-R2 and R2-B) is the same. What is the length of the three links (in km)? Show your work. 5 of 10 2. [10pt] Consider the following DNS example with requesting host (cis.poly.edu) looking to resolve gaia.cs.umass.edu. IP address of each server and host is given in the diagram. Answer if the following is possible or not possible, and explain in 1 sentence why it is possible or not possible. a. The local DNS server directly contacts the .edu TLD server and skips steps 2 and 3. b. An RR included in the response message 7 is of type NS. 6 of 10 c. There are two RRs included in response message 5, one of type A and another with type NS. d. The type A RR included in the response message 3 is an authoritative response. 3. [15pt] Consider a cross-country link between node A and node B. Node A is sending data to Node B using a pipelined protocol where the size of A’s window is n packets. The RTT is 190ms and both nodes are connected by a channel with a transmission rate of 800 kbps. Suppose the packet size is 1000 bytes including header and data. a. What is the window size (value of n) that has to be for the channel utilization to be 75 percent? Utilization of the channel is the fraction of time the sender is actually busy sending bits into the channel. Assume that the receiver has an infinite buffer. Explain your answer. 7 of 10 b. Now assume that node A uses flow control to make sure it does not overflow the available buffer at node B. The size of the maximum available buffer at node B is 10KB. Considering the flow control and your answer in (a) above, what can be the maximum window size (value of n)? Explain your answer. 4. [10pt] Consider the GBN protocol with a sender window size of 5 and sequence number range of 1024. Suppose at time t, the next in-order packet that the receiver is expecting has the sequence number 40. Assume that the medium does not reorder messages. Answer if the following is possible or not possible, and explain in 2 sentences why it is possible or not possible. a. Sender’s window is [40, 44]. b. Sender’s window is [35, 39]. 8 of 10 c. Packet 33 is unacknowledged (is in sender’s current window). d. ACK for packet 38 is sent by the receiver and is currently propagating back to the sender. e. Receiver has sent an ACK for packet 41. 5. [15pt] Compare GBN, SR, and TCP (no 500ms delayed ACK). Assume that the timeout values for all three protocols are sufficiently long such that 5 consecutive data segments (Seq. no. 1, 2, 3, 4 and 5) and their corresponding ACKs can be received (if not lost in the channel) by the receiving host (Host B) and the sending host (Host A) respectively. Suppose Host A sends 5 data segments to Host B, and the 2nd and 4th segments (sent from A) are lost. In the end, all 5 data segments have been correctly received by Host B. a. How many segments has Host A sent in total? What are their sequence numbers (in order)? Answer this question for all three protocols. 9 of 10 b. How many ACKs has Host B sent in total? What are their sequence numbers (in order)? Answer this question for all three protocols. 6. [10pt] Consider the following figure of TCP Reno’s operation. a. For each increase, decrease or transition, identify the name of the phase, event that causes a change/transition and value of ssthresh. 10 of 10 b. Why is the window size increased linearly starting transmission round 6? Of course, this is because ssthresh is reached, but why the designers of TCP thought of linear increase beyond this point and not exponential increase? Answer in no more than 2 sentences. c. At transmission round 16, the window size is decreased but not exactly in half. Why is it not exactly decreased to half at transmission round 17? Answer in no more than 2 sentences. d. Is the network congestion more at transmission round 16 or transmission round 36? How do you know that? Answer in no more than 2 sentences.
欢迎咨询51作业君
