cs438_hw3

.pdf

School

University of Illinois, Urbana Champaign *

*We aren’t endorsed by this school

Course

438

Subject

Electrical Engineering

Date

Apr 3, 2024

Type

pdf

Pages

Uploaded by willyrom

CS/ECE 438: Communication Networks Spring 2024 Homework 3 Handed Out: March 19 th , 2024 Due: 11:59pm, March 24 th , 2024 TA: Federico Cifuentes-Urtubey • Homework assignments must be submitted online through GradeScope . We will not correct anything we do not understand. Contact the TAs via Ed if you face technical difficulties in submitting the assignment. Note: Please select your page(s) for each question during submission on GradeScope; Otherwise, your submission will not be graded. • Homework assignments can be done in groups of two , but only one person needs to submit on Grade- Scope. Remember to include your partners name in the solution and on GradeScope by editing "Group Members" . It is your responsibility that your partner’s name is included. For detailed instruction please refer to ( https://youtu.be/rue7p_kATLA ). • You can use Ed to find a partner. We highly recommend working in groups. You will not get extra credit for working alone. • Please use Ed and come to office hours if you have questions about the homework. Failure to under- stand the solutions will be the student’s fault. • While we encourage discussion within and outside of the class, cheating and copying is strictly pro- hibited. Copied solutions will result in the entire assignment being discarded from grading at the very least and a report filed in the FAIR system. It is also your responsibility to ensure that your partner obeys the academic integrity rules as well. 1 TCP RTT Estimation – 7 points One difficulty with the original TCP SRTT estimator is the choice of an initial value. In the absence of any special knowledge of network conditions, the typical approach is to pick an arbitrary value, such as 3 seconds, and hope this will converge quickly to an accurate value. If this estimate is too small, TCP will perform unnecessary retransmissions. If it is too large, TCP will wait a long time before retransmitting if the first segment is lost. Also, the convergence might be slow. 1. Choose α = 0 . 4 and RTT-timeout (0) = 1 seconds, and assume all measured RTT values = 0.5 second with no packet loss. What is RTT-timeout (20) ? Recall, RTT-timeout ( k + 1) = α × RTT-timeout ( k ) + (1 − α ) × RTT ( k + 1) Describe your solution approach AND provide the numerical result (approximate to 4 th decimal place). 2. Using the same values as in above part, what happens if we use α = 0 . 2 or α = 0 . 95 ? Provide a numerical result for RTT-timeout (20) in both cases, then describe the effect of a larger or smaller α on the RTT estimation procedure. 1

2 Fair Queuing – 18 points Consider Fig. 1. Now consider the following scenarios: Figure 1: • Packets are scheduled using FIFO policy • Packets are scheduled using Highest Priority First policy. Assume odd-numbered pack- ets are high priority and even-numbered packets are low priority. • Packets are scheduled using Round Robin policy. Assume that packets (1, 4, 5, 7, 8, 9, 10) are from class 1, and packets (2, 3, 6, 11, 12) are from class 2. Scheduling starts with class 1 at t = 0 . • Packets are scheduled using Weighted Fair Queuing (WFQ) policy. Assume there are three classes. Let us denote the packet ID of packet i as ID i . Class j (where j = 0 , 1 , 2 ), will contain packets which satisfy i | ID i %3 = j . Let the three classes (i.e. Class 0, 1 and 2) have the weights 3, 2 and 1 respectively. Scheduling starts with class 0 at t = 0 . Note : In case of ties under the above schemes, you must resort to FIFO scheme. In Fig. 1, you can assume that Packet 1 arrives before Packet 2 at t = 0 , Packet 3 before Packet 4 at t = 1 and so on. Also assume that packets can be immediately scheduled for transmission when they arrive. For Round Robin and WFQ, you must skip a Class if packets are not available for that particular class. 1. Fill this Table. 1 for which packet departs at each time point and the delay of each packet (i.e., delay is time interval between the arrival and departure of a packet). 2. What is average delay for these 4 polices? For Highest priority, Round Robin, and WFQ, you must also list average delay for each class separately. (In WFQ, the average delay among different classes might not be perfect in this question.) 3. What observations can you draw about the average delay from the above 4 polices? Two concise observations are sufficient. 2

Table 1: Packet Scheduling Time of Departure (t in sec) FIFO Highest Priority Round Robin WFQ Packet Delay Packet Delay Packet Delay Packet Delay 1 1 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 10 11 12 3 Forwarding and CIDR – 15 points 1. Consider a router that interconnects three subnets: Subnet A, Subnet B, and Subnet C. Suppose all of the interfaces in each of these three subnets are required to have the prefix 200.64.16.0/24. Also suppose that Subnet A is required to support up to 80 interfaces, and Subnets B and C are each required to support up to 30 interfaces. Provide three network addresses (of the form a.b.c.d/x) for each of the above subnet that satisfy these constraints. Suppose a router has built up the routing table shown below in Table. 2. CIDR addresses are used, with "/22" indicating a mask of 22 1’s followed by 10 0’s. Net/Masklength NextHop 220.174.252.0/22 R1 220.174.240.16/29 Interface 1 220.174.240.10 R2 220.174.240.0/20 Interface 3 220.174.240.128/25 R3 220.174.248.0/22 Interface 2 default Interface 4 Table 2: 2. How many individual IP addresses match each Net/Masklength pair? (Compute this for all entries in the table except the last one, i.e. for the default Masklength entry). 3. The router can deliver packets directly over interfaces 1, 2, 3, 4, or it can forward to routers R1, R2, R3. Specify the next hop for each of the following destinations. Use the longest prefix match, i.e., if a destination matches more than one line of the table, 3

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version