This exercise will introduce you to the Open Network Laboratory (ONL). Topics include basic packet routing, and traffic monitoring. You should read through the entire exercise first to familiarize yourself with the exercise.

Getting Started

• Familiarize yourself with the main ONL web page ( http://onl.arl.wustl.edu).
• Skip this step if you plan to use the pre-assigned account names.
  Get an ONL account by following the Get an account link and following the instructions.
• After getting an account, read the Getting Started page and follow the directions on getting the RLI.jar file and creating an SSH tunnel to the ONL Central Resource Daemon.
• Read the following tutorial sections
  • Overview (in detail)
  • The ONL Architecture (scan)
  • The Remote Laboratory Interface (in detail)
• Get this ONL configuration file (cfg1-1.exp) and store it in the file system of the host where you plan to run the RLI. I usually put it in the same directory where my RLI.jar file is stored.

Exploring a Simple Pre-Configured Experiment

In this part of the exercise, you will be asked to familiarize yourself with the experiment file that you got above and try to determine the value of some configuration parameters. You will NOT connect to the ONL testbed until the following section. But you will be asked to determine various configuration parameter values by opening various RLI menus.

Start the RLI, open the cfg1-1.exp configuration file, and answer the following questions by viewing the appropriate menus. Hint: You can determine parameters values by using the left or right mouse buttons and selecting appropriate icons in the RLI.

Questions:
Consider a packet P that enters port 2 with a destination IP address of D where D is specified below.

1. Which route entry will be used at port 2 if D = 192.168.1.35? Explain.
2. Which route entry will be used at port 2 if D = 192.168.1.64? Explain.
3. For D = 192.168.1.64, indicate the path (i.e., sequence of ingress ports, egress ports and host interfaces) taken through the router once P enters ingress port 2. Explain.
4. What changes to the routing table at port 2 would be needed to forward packets to port 0 when D = 192.168.1.X where X is between 118 and 121 inclusive. Explain.
5. What is the capacity (Mbps) of the egress link at port 3? Port 4? Hint: See Tutorial => Filters, Queues and Bandwidth => Modifying Link Rate
6. Why are there eight (8) VOQs at each input port?
7. Suppose that we have a steady stream of UDP packets entering port 2 that are destined for port 4. At what input rate (Mbps) would a queue begin to build at VOQ(2,4)? Explain.
8. What is the meaning of the threshold parameter in Queue Table entries?

Running a Simple Pre-Configured Experiment

In this part of the exercise, you will run traffic experiments on a network that has been configured through the RLI using the cfg1-1.exp configuration file. You must have a resource reservation to do this.

In preparation for the experiments in this section, you should know how to run the ping command with and without the -c flag and the netstat command with -i flag. If you are unfamiliar with these command, read the man pages on these two commands on onl03.arl.wustl.edu (or any Linux host) using the commands:

          man ping
          man netstat

Follow the instructions below:

Instructions:

• Log into the ONL web site, and reserve 1 cluster and 0 extra hosts for 1 hour.
• Create a tunnel to the ONL Central Resource Daemon by opening an SSH connection to onl03.arl.wustl.edu with a tunnel to port 7070.
• Start the RLI; open the cfg1-1.exp configuration file; and then, get testbed resources by selecting File => Commit.
• After the resources have been successfully allocated to your experiment, answer Question 9. Hint: You can determine parameters values by using the left or right mouse buttons and selecting appropriate icons in the RLI.
• SSH into the host corresponding to n1p2 and run the following commands:

            netstat -i
            ping n1p3 -c 10
            netstat -i
  

  Note 1: You can only SSH into onl03.arl.wustl.edu when outside of the ONL testbed. This means that you will need to SSH first into onl03 and then into your desired host.
  Note 2: You should SSH into the ethernet interface, not the ATM one; i.e., the interface will NOT be n1p2. You will be SSHing into something like onl02, onl19, onl15 or onl13. (Do you know how to determine the correct name???)
  
Answer Questions 10 and 11.
• Modify the last entry in the route table at port 2 so that its prefix/mask entry is 192.168.1.64/16 and the next hop is 3. (Remember to Commit.) Repeat the previous experiment. Answer Question 12.

Questions:

9. The host labeled n1p2 actually has two network interfaces: one ATM interface connected to router port 2 and one ethernet interface connected to the control network (not shown by the RLI). The label n1p2 actually refers to the ATM interface and has an IP address of 192.168.1.48. What is the name of the ethernet interface?
10. The output of the netstat command should show the number of packets transmitted (TX-OK) and the number of packets received (RX-OK) for the atm0 interface. Explain why the changes in the TX-OK and RX-OK columns for the atm0 interface make sense.
11. There are three plots on your traffic graph. Explain why the plots make sense given that you just ran the ping command. Hint: Pass your mouse over each of the three labels (e.g., IPPBW 2) to highlight the corresponding line.
12. How has the Route Table change changed the traffic plots? Explain why the plots make sense.

Creating and Testing Your Own Configuration File

In the previous section, you started with an existing configuration file. In this section, you will create, save and test your own configuration file.

Follow the instructions below:

Instructions:

• If you have not exited the RLI, do so now by selecting File => Exit.
• Start the RLI again, but do NOT open a configuration file. Begin by adding a cluster (File => Topology => Add Cluster). At this point, all settings have default values, and all ports are identically configured. This means that you can determine the settings of all ports by examining any single port (e.g., port 2). Answer Questions 13 and 14.
• Setup routing such that any packet coming from the host(s) attached to ports 1, 2 or 3 first go to port 4, out onto a loopback link attached to port 5, into port 5, then to the port attached to the destination host. For example, a packet from host n1p1a destined for n1p2 should go through the GigE switch, into port 1, out port 4 onto a loopback link to port 5, then to port 2, and finally to n1p2. Answer Question 15.
• Save the current configuration to the file cfg1-1a.exp.
• Add a monitoring panel that will display three bandwidth plots: 1) bandwidth coming into port 3; 2) bandwidth coming into port 5; and 3) bandwidth going out of port 2.
• Add a second monitoring panel called path 2-4-3 that will display three bandwidth plots: 1) bandwidth coming into port 2; 2) bandwidth coming into port 4; and 3) bandwidth going out of port 3.
• Ping n1p2 from n1p3. Answer Question 16.

Questions:

13. What are the default values of the following parameters:
    • Routing Table
    • Switch Rate
    • Link Rate
    • Thresholds and rates of VOQs
    • Thresholds and rates of Datagram Queues
14. What are the contents of the routing tables at ports 1-7?
15. What are the contents of the routing tables at ports 1-5?
16. Submit screen shots of the traffic displays, and explain why these displays demonstrate that your routing tables are correct.