3 Dynamic Routing A (RIP and OSPF)

Transcription

1 אוניברסיטת בן גוריון בנגב המחלקה להנדסת מערכות תקשורת רשתות תקשורת מחשבים - 2 קורס מעבדה בתקשורת מחשבים 3 Dynamic Routing A (RIP and OSPF) בשבוע (29/11/2011) 5 בשבוע (13/12/2011) 7 מעבדה מספר שם המעבדה תאריך הגשת דוח מכין תאריך הגשת דוח מסכם 1. הוראות כלליות דוח מכין יורכב מתשובות לשאלות הכנה. דוח מסכם יורכב מתשובות לשאלות בגוף מעבדה, הצגת נתונים וניתוחם. דוח מסכם יהיה קובץ יחיד בפורמט word או.pdf הנתונים הרלוונטיים מקבצים ששמרתם צריכים להיות מועתקים לקובץ הדוח. אין לצרף קבצים נוספים. כל כתובות ה- IP שתקנפגו במהלך המעבדה יהיו כפי שיוגדר בספר או בקובץ ההנחיות ועם השינוי הבא: האוקטט השני ("octec1.octet2.octet3.octet4") של כל כתובת IP יוחלף במספר, X אשר יהיה פונקציה +2 = X, כאשר - A שלוש ספרות אחרונות של ת.ז. של בן הזוג ( A+ של תעודות הזהות שלכם: (B mod 252 הראשון, ו- - B שלוש ספרות אחרונות של ת.ז. של בן הזוג השני. יש לרשום את מספר ה- X שלכם בברור בתחילת כל דוח מסכם. הגשת כל הדוחות היא חובה, הגשה בזוגות. הגשת דוח מכין בתחילת המעבדה, אין אפשרות לעשות את המעבדה ללא דוח מכין. כל דוח מסכם יכיל עמוד שער עם שם המעבדה ושמות הסטודנטים. הוסיפו במקומות הנחוצים בדוח מידע מקבצים ששמרתם. המידע צריך לתמוך בתשובה שניתנה. הוסיפו מתוך הקבצים אך ורק את המידע הרלוונטי. הבודק הונחה להוריד נקודות על כל מידע לא רלוונטי וארוך מידי חומר קריאה פרוטוקול :RIP "ספירה לאינסוף" ב- :RIP פרוטוקול :OSPF תקראו את מהלך מעבדה בקובץ הנוכחי (במעבדה הזאת כל ההוראות לביצוע מופיעות בקובץ הנוכחי ללא הפניות לספר), תלמדו את הפקודות שאתם לא מכירים.

2 3. שאלות הכנה 1. What are the main differences between a distance vector routing protocol and a link state protocol? Give examples of each type of a protocol. 2. What is the main difference between RIP version 1 and RIP version 2? 3. Explain what it means to run RIP in the passive mode. 4. Explain the meaning of triggered updates in RIP. 5. Explain the count-to-infinity problem in RIP. Give and explain an example where the count-to-infinity loop contains at least 3 routers. 6. Which transport service is used to deliver RIP messages? 7. What message types (commands) does RIP define? Shortly explain the role of each message and in what cases it is sent. 8. Which transport service is used to deliver OSPF messages? 9. What are the OSPF message types? What is the role of each message? 10. Explain how OSPF supports big autonomous systems by using several route databases instead of one big database. What kind of routing does this scheme define?

3 4. סעיפים לביצוע RIP Part RIP Part A Configuring RIP Network Setup: lab3 a 1. Turn off network interfaces "eth0" and "eth1" of all PCs and Routers (later you will be instructed to configure only the interfaces needed for this lab): ifconfig eth0 down ifconfig eth1 down 2. In order to give PCs an ability to participate in dynamic routing algorithms, all PCs are equipped with Quagga program (the same as Routers). Reset Quagga on PCs and Routers (this is needed to reset all routing settings and to stop all dynamic routing algorithms issued by other students via Quagga): /etc/init.d/quagga restart

4 3. Set up the topology lab3-a. Configure PC's and Routers' IP addresses according to the topology lab3-a. (Don t forget to Enable IP forwarding on routers.) 4. Enable RIP on Router1, Router2 and Router3 a. router1:~# telnet b. ripd> enable c. ripd# configure terminal d. ripd(config)# router rip e. ripd(config-router)# version 2 f. ripd(config-router)# network 10.x.0.0/16 g. ripd(config-router)# end 5. Enable RIP on all the PCs: a. pc1:~# telnet b. ripd> enable c. ripd# configure terminal d. ripd(config)# router rip e. ripd(config-router)# version 2 f. ripd(config-router)# network 10.x.0.0/16 g. ripd(config-router)# passive-interface eth0 h. ripd(config-router)# end 6. Once the routing tables have converged: a. save routing tables on all Routers using show ip rip in quagga - privileged mode. b. save routing tables on all PCs using the route en Linux command. 7. In your lab report: a. How long (approximately) does it take for all routing tables to converge? How do you know that the routing tables have converged? b. Discuss the differences and similarities in the output of the commands: show ip rip and route en. c. Explain the meaning and purpose of each of command entered into Quagga at stages 4 and 5. d. Explain why passive-interface command was entered on the PCs and not on the Routers. 8. You should now be able to ping any pc from any other pc or router. Make sure this is indeed so. a. Display the route from pc1 to pc4 with traceroute and save the results. 9. Start Wireshark on PC1, PC2, PC3 and PC4 and set a capture or a display filter to RIP packets. Answer the following questions in your lab report, for each include captured packets to support your answers. a. What is the destination IP address on the RIP packets? b. Do routers forward RIP packets? Does PC1 receive RIP packets sent by Router3? c. Which type of routing message do you observe (the type is indicated by value of the field command)? For each packet that you observe, explain the role of that message in the RIP protocol. d. A RIP message may contain multiple routing table entries. How many bytes are consumed in a RIP message for each routing table entry? e. What information is transmitted for each entry? Use a single RIP packet to explain the fields in a RIP message.

5 10. Stop the Wireshark and save the relevant data (if necessary, select the print details option). RIP Part B Good News Propagation 11. Now we are going to find out how fast good news propagate via RIP. 12. Just before we add an extra router (Router 4) to our network, let s first prepare it. Configure Router 4 IP addresses according to the topology lab3-b and enable IP forwarding. 13. Now, we are going to connect Router 4 to our network. Please set up the following topology: Network Setup: lab3 b 14. Configure the RIP protocol for Router 4 (as in step 4). 15. Observe as the routing tables are updating, and after the convergence save the output of route en on all the PCs. 16. In the lab report: a. Include the routing tables before the topology change (step 6.b) and after router 4 has been added and the algorithm has converged (step 15). b. Discuss the time it took the algorithm to converge.

6 RIP Part C Bad News Propagation 17. Now we will find how fast (or slow) bad news (link failures) propagate via RIP. 18. Issue an infinite ping to PC1 from PC4 while saving the output to file: ping 10.x.1.10 tee somefile.txt. You should see successful pings. Do not terminate the command until you are instructed otherwise. 19. Now we want to simulate a link failure. In order to do that, restart Quagga on Router4. By doing so, we will clear Router 4 routing table from dynamic routing entries and prevent it to participate in the RIP algorithm. 20. Wait until the ping command is successful again. That will happen when an alternate path is found and may take several minutes. 21. Stop the ping command with ctrl+c and count the number of lost packets the time it took RIP to update the routing tables. 22. Save the ping statistics output (what appears at the bottom of the terminal). 23. Calculate the time it took RIP to converge.

7 RIP Part D Count to Infinity Problem Network Setup: lab3 c 24. Set up the topology lab3-c. 25. Configure IP addresses and enable IP forwarding on PC3 (according to lab3-c). 26. Restart Quagga on PC3 and configure RIP protocol as was done for the routers (step 4). 27. Re-configure RIP on router 4. This should be done because Quagga was restarted. 28. On PC3, start capturing traffic with wireshark on both eth0 and eth1. Set a filter to display only RIP packets. 29. On PC1 and PC4 use netstat -rn to observe that the routing tables are updated accordingly. 30. Make the routes to/from network 10.x.1.0/16 through Router1 unattractive: a. Connect to Quagga in the Router1/RIP and enter the config-router mode. b. Execute: ripd(config-router)# redistribute connected metric Wait until the routing tables on all the routers have converged. Since the cost of the interface on Router1 is set high, you should observe that the traffic from all hosts and routers to network 10.x.1.0/24 passes through PC3. a. This can be verified by traceroute (from Linux) PC Issue a continuous ping command from PC2 to PC1 (without c).

8 33. On PC3, disable eth1 to break the connection between PC3 and the network 10.x.1.0/24. PC3 will send out a RIP packet indicating that the cost metric to network 10.x.1.0/24 is infinity (16). When the interface is disabled, the pings from PC2 to PC1 will fail. The ping command will be successful again as the routing tables will converge. a. Notice that Quagga has a built-in mechanism that makes it difficult to simulate the count-to-infinity problem. This mechanism is called triggered updates. When PC3 feels that its link went down, it immediately sends this information to its neighbors and thus, this information almost immediately distributes over the network. 34. Now you should observe a slow convergence due to the count-to-infinity problem by observing the RIP messages in the Wireshark. a. Probably, you didn t see the count-to-infinity problem due to the triggered updates mechanism. Please describe in your report scenario in which you would observe the count-to-infinity problem despite the triggered updates mechanism. (Hint: think about timing of regular RIP update messages and triggered update messages). b. Wait until the ping from step 32 is successful again. c. Save the RIP packets that were captured by Wireshark (print them selecting print detail option). Save only the packets that are needed to explain the count-toinfinity problem (if you didn t observe the count-to-infinity, just explain what happened in the network using the captured packets).

9 OSPF Part Network Setup: lab3 d

10 OSPF Part A Configuring OSPF 1. Configure topology lab3 d on your netlab. 2. Turn off eth0 and eth1 on all PCs and routers. This is necessary to allow the detection of the LAN in the routing table (which is usually messed-up after the lab). 3. Configure the IP addresses of all devices according to the drawing. 4. Enable IP forwarding on all routers and all PCs. 5. Reset Quagga on PCs and Routers by issuing the following command: /etc/init.d/quagga restart This step is necessary to cancel any changes other students have made to Quagga's configuration. 6. Now configure OSPF on all routers and PCs: I. Configure OSPF for PC1: PC1% telnet Password: zebra ospfd> enable ospfd# configure terminal ospfd(config)# router ospf ospfd(config-router)# network 10.x.0.0/16 area 1 ospfd(config-router)# router-id 10.x.1.1 ("router-id" should be set to the IP address of eth0). ospfd(config-router)# no passive-interface eth0 ospfd(config-router)# no passive-interface eth1 ospfd(config-router)# end ospfd# exit II. Notice, that the value for "router-id" is different for each device In your lab report, explain the meaning and purpose of each of the commands entered into Quagga at this stage. III. Similarly, configure the rest of the PCs and routers. IV. When the OSPF configuration is complete, all hosts and routers should be able to communicate with one another. Check the correctness of the configuration using traceroute and ping commands.

11 OSPF Part B Convergence of OSPF 7. You will now observe the interactions of OSPF nodes after a change in network topology. I. On PC1 start capturing traffic on eth0 using wireshark. Use a display filter to show only OSPF packets. II. From PC3 run a traceroute to PC4. From the output, determine whether the path from PC3 to PC4 includes Router 3 or Router 4. Include the result of the traceroute along with your explanation of it in your lab report. III. Issue a ping command from PC3 to PC4, do not specify the number of pings to send. Instead, send pings continuously until told to stop. Use the Linux tee command to save the command output while also seeing it on the screen. IV. Check whether the path from PC3 to PC4 includes Router3 or Router4: i. If the path includes Router3, perform ifconfig eth1 down on Router3. ii. If the path includes Router4, perform ifconfig eth1 down on Router4. V. Now, OSPF updates the routing tables. Observe the process from the wireshark window on PC1. i. How quickly are OSPF messages sent after the interface is turned off? ii. How many OSPF messages are sent? iii. Which type of OSPF packets is used to flooding link state information? iv. Describe the flooding of LSAs to all routers v. Which type of encapsulation is used for OSPF packets (TCP, UDP or other)? vi. What is the destination address of OSPF packets? VI. Wait until the ping command issued from PC3 is successful again, that is until ICMP echo replies arrive at PC3. This happens when the routing tables have been updated. VII. Stop the ping command using Ctrl+C. Using the ping output count the number of lost packets and calculate the time it took OSPF to update the routing tables. Note that pings are sent once a second. VIII. Issue another traceroute command from PC3 to PC4. Save the output and note the new route to PC4. IX. Save the link state database on all routers and PCs and verify that all routers indeed have the same link state database. Use the "show ip ospf database" to see Quagga's OSPF database. i. Compare the output of the command on all PCs and routers. Can you confirm that the link state databases are identical? ii. Include the link state database of PC2 in your lab report. X. Stop Wireshark on PC1 and save the different types of OSPF packets captured. Save one copy of each OSPF packet IN DETAIL (including all the packet bytes). i. Include one example of each packet in your lab report. ii. Pick a single link state advertisement packet captured by wireshark and describe how to interpret the information contained in the link state advertisement.

12 OSPF Part C Hierarchical Routing in OSPF

13 8. In OSPF, networks can be divided into areas, to significantly reduce the amount of topological information routers have to learn. In OSPF all areas must be connected to Area 0 which is known as the backbone area. You will now define Hierarchical OSPF with two areas (area 1 and 2) connected through a backbone area (area 0). Routers that connect two areas are called area border routers. I. Restart Quagga on all PCs and Routers using /etc/init.d/quagga restart II. Don t forget to enable the interface that was disabled in 7.IV. III. Start wireshark on eth0 of PC1 IV. Configure Quagga on all PCs and Routers by executing the following commands: i. telnet ii. ospfd> enable iii. ospfd# configure terminal iv. ospfd(config)# routetr ospf v. ospfd(config-router)# router-id 10.x.1.1 ("router-id" should be set to the IP address of eth0 of the device) vi. ospfd(config-router)# network 10.x.1.0/24 area 0 (Set appropriate network and area) vii. ospfd(config-router)# network 10.x.2.0/24 area 1 (Set appropriate network and area) viii. ospfd(config-router)# no passive-interface eth0 ix. ospfd(config-router)# no passive-interface eth1 x. ospfd(config-router)# end V. In your lab report, explain the meaning of these commands. VI. Save the link state database of all routers and PCs. You can see the link state database by typing show ip ospf database in Quagga s privileged mode. i. Attach the output to your lab report. ii. Compare the link state databases to those saved in part 6. Discuss the differences. iii. Which information do routers in Area 1 have about Area 2? Which information do they have about the backbone area? iv. How much information do the routers in the backbone area have about the topology of Area 1 and Area 2? v. How do the IP routers in Area 1 know how to forward traffic to Area 2? VII. Display the area routers known to Router1 from Area1, with the command Router1# show ip ospf border-routers, include the output in your lab report, explain it s meaning. VIII. Save the wireshark output of OSPF packet types (with the Print detail option) that you didn t observe in part 6. בהצלחה!