COM-208: Computer Networks - Homework 1

Transcription

1 COM-208: Computer Networks - Homework 1 1. Design an application-level protocol to be used between an TM (automatic teller machine) and a bank s centralized server. The protocol should allow: verifying a user s card and password querying the account balance (which is maintained at the server) making a money withdrawal The protocol must be able to handle the case in which there is not enough money in the account to cover the withdrawal. Specify the protocol by listing the messages exchanged and the actions taken by the TM or the bank s server on transmission and receipt of messages. Sketch the operation of your protocol for the case of a simple withdrawal with no errors, using a timing diagram as seen in the lecture. Then sketch the operation in case there are not sufficient funds to complete the withdrawal. Host Host R,l,c Figure 1: Two hosts connected by a single link 2. Consider two hosts, and, connected by a single link of rate R bps as shown in Figure 1. The link s cable length is l and the signal propagation speed along the cable is c. Host sends to two s back-to-back, the first having size L 1 and the second L 2. Suppose c = m/s, R = 100 Mbps, l = 200 km, L 1 = 1 kb and L 2 = 10 kb. a. Express the propagation delay on the link, d prop, in terms of l and c. b. Determine the transmission delay of each, d trans,1 and d trans,2, in terms of L 1, L 2 and R. c. What is the queuing delay of each, d queue,1 and d queue,2? d. Draw a timing diagram showing the transmission of each and the delays (d prop, d trans, d queue ). e. Compute the transfer time, i.e. the time elapsed between the moment host starts transmission to the moment host receives the last bit of the data. Note: in computer networks, we use the following measurement units related to data transfer: bps = b/s = bits per second ps = /s = bytes per second, with 1 byte = 8 bits 1 kbps = 1000 bps (not 1024!), 1 Mbps = 1000 kbps = 10 6 bps, 1 Gbps = 1000 Mbps = 10 9 bps etc. Delays are often specified in the unit that yields the most compact representation, for example 50 ms instead of 0.05 s. 3. Consider again the scenario in the previous problem, except the s are now sent over a path consisting of multiple links, connected by relays, as shown in Figure 2. The transmission method of the relays is store-and-forward, and their processing delay is d proc = 1 us for each. Consider that each link has the same properties as in the previous problem, i.e. transmission rate R, length l and signal propagation speed c. What is the transfer time if the number of links is: 1

2 Host Relay R 1 Relay R 2 Relay R N-1 Host R,l,c R,l,c R,l,c Figure 2: Two hosts connected by a path with multiple links with identical properties a. 2 b. 3 c. N Draw a timing diagram to help justify your answer. 4. Consider the scenario in the previous problem, except the relays now use circuit ing, i.e. they do not store-and-forward s but instead immediately transmit each bit they receive before waiting for the entire to arrive. What is the transfer time if the number of links is: a. 2 b. 3 c. N Draw a timing diagram to help justify your answer. 5. (P4) Consider the circuit-ed network in Figure 3. There are 4 internal links, each supporting up to 4 connections. ny connection must pass through an internal link (it cannot enter and exit the network at the same ). a. What is the maximum number of simultaneous connections that can be in progress? b. Suppose that all connections enter or exit the network through es and C. What is the maximum number of simultaneous connections that can be in progress? c. Suppose we want to make four connections that enter or exit the network through es and C, and another four connections that enter or exit the network through es and D. Can we simultaneously route all eight connections through the network? Switch Switch Switch D Switch C Figure 3: circuit-ed network 2

3 6. (P8) Packet ing vs circuit ing Multiple users share a 3 Mbps link. Suppose each user requires 150 kbps when transmitting, but each user transmits only 10% of the time. Note: Subquestions (c) and (d) of this problem are optional, as they require special knowledge about probability theory, that you may have not yet acquired. Yet, they are given here because they show how the probabilities around statistical multiplexing in ing (slide 50 of Lecture 1) are computed. a. For the beginning of the problem, assume that circuit ing is used. In particular, a circuit in the link is implemented with Frequency Division Multiplexing (FDM) or Time Division Multiplexing (TDM). How many users can be supported? Does this number differ when FDM is used vs TDM? b. For the remainder of this problem, suppose that ing is used. Find the probability p that a given user is transmitting at any time. c. (Optional) Suppose that there are 120 users. Find the probability that at any given time, exactly n users are transmitting simultaneously. (Hint: use the binomial distribution.) d. (Optional) Find the probability that there are 21 or more users transmitting simultaneously. e. Which method provides better performance guarantees? Which method uses more efficiently the available resourses (link bandwidth)? Explain. 7. (P31) In modern -ed networks, including the Internet, the source host segments long, application-layer messages (for example, an image or a music file) into smaller s and sends each separately into the network. The receiver then reassembles the s back into the original message. We refer to this process as message segmentation. Figures 4 and 5 illustrate the end-to-end transport of a message with and without message segmentation. message source destination Figure 4: Sending a message without segmentation source destination Figure 5: Sending a message with segmentation Consider a message that is bits long. Each of the 3 links in the network has the rate 2 Mbps. Ignore propagation, queuing, and processing delays. a. Consider sending the message without message segmentation. How long does it take to transfer the message from to the first? Keeping in mind that each uses store-and-forward ing, what is the total time to transfer the message from to? b. Now suppose that the message is segmented into 800 s, with each being 10,000 bits long. How long does it take to move the file from to? Discuss. c. In addition to reducing delay, what are other reasons to use message segmentation? d. Discuss the drawbacks of message segmentation. 8. In Problem 3, you computed the total transfer time of two s over N links. In this problem, you will generalize your results for multiple s. Consider Figure 2, and assume that: host sends to, P s back-to-back, each of which has length L; the number of link is N and each link has transmission rate R; 3

4 any processing and propagation delays may be ignored, as in the previous exercise (message segmentation). there is no other traffic in the network (i.e. there is no queuing delay due to other s) nswer to the following questions: a. What will be the end-to-end delay for sending P s from to? b. What will be the end-to-end delay if the s do not have the same length, i.e. the s lengths are L 1, L 2,..., L P? 9. (P21) Consider Figure 6. R 1 2 R 1 1 R 1 N R 2 1 R 2 2 R 2 N R M 1 R M N R M 2 Figure 6: network topology with multiple paths This network consists of M paths between the server and the client. No two paths share any link. Each path k (k = 1,..., M) consists of N links, with transmission rate R k 1, R k 2,..., R k N. a. If the server can only use path 1 to send data to the client, what is the maximum throughput that the server can achieve? b. If the server can use all M paths simultaneously to send data, what is the maximum throughput that the server can achieve? Rs Rc server client Figure 7: Two links with different transmission rates. 10. (P23) Consider Figure 7. ssume that we know the bottleneck link along the path from the server to the client is the first link with R S bits/sec. Suppose we send a pair of s back to back from the server to the client, and there is no other traffic on this path. ssume each of size L bits, and both links have the same propagation delay d prop. a. What is the inter-arrival time at the destination? That is, how much time elapses from when the last bit of the first arrives until the last bit of the second arrives? b. Now assume that the second link is the bottleneck link (i.e., R C < R S ). Is it possible that the second queues at the input queue of the second link? Explain. Furthermore, suppose that the server sends the second T seconds after sending the first. How large must T be to ensure no queuing before the second link? Explain. 11. (P24) You need to transfer urgently 300 T data from Lausanne to London. You have available a 1 Gbps dedicated link for data transfer; alternatively, you can send the hard drives by post with 48-hour delivery. Is it faster to transmit the data via this link or instead use the post? Explain. 12. (P25) Two hosts, and, are separated by 20,000 kilometers and are connected by a direct link of R = 2 Mbps. The signal propagation speed over the link is meters/sec. a. Calculate the bandwidth-delay product, R d prop. b. Consider sending a large data stream from Host to Host. What is the maximum number of bits that will be in transit on the link at any given time? 4

5 c. Visualize the link as a pipe filled with a sequence of bits. What is the length (in meters) of a bit in the link, i.e., what fraction of the link corresponds to one bit? d. Express the length of a bit in terms of the propagation speed s, the transmission rate R, and the length of the link m (are all needed?). 13. (Open question) To reduce the complexity of the system, the Internet is based on a set of layered protocols and the payload is encapsulateddecapsulated whenever it is transferred from one layer to another. Think of another real-world system that is structured in layers and uses encapsulation/decapsulation whenever the payload is transferred from one layer to another. Describe the layers of the system and the encapsulation-decapsulation process. (Hint: a postal service) 14. (Open question) Imagine yourself being the engineer who is responsible for the security part in the following applications. Which type of security attack would you prefer to avoid in each case? Explain why and provide a high-level systematic solution to address the attack. a. n online banking system which should enable customers use their private accounts and perform financial transactions (just like TM banking). b. weather website which should be available 24 hours per day. c. chat application (similar to Snapchat) which should provide ephemeral and private messaging to assure senders that their messages are burnt after reading. I.e. once the recipient of a message has viewed its content, the message is automatically deleted. The recipient may make a message copy, but in that case the sender is explicitly notified. 5