Objective: The objective of this exercise is to better understand the effects of latency and throughput.

Preparation: Login to a computer and read the man page on ping. Ping sends an ICMP ECHO_REQUEST packet to network hosts. Several options allow you to set the recipient host, packet size, number of packets, among other things. You may be interested in reading about The Story of the PING Program.

Ping measures round trip time (RTT). The measured RTT between computers will vary depending on queueing delays that may be encountered. In general,

Exercise: After learning how to use ping, do the following:

1. For message sizes of 8, 50, 100, 200, 500, 1000, 2000, 4000, 6000, and 8000 do the following:
   From your login machine, ping the machine that you think is closest to your machine on the network. If you don't know what machine that might be, ping comp. (If you're logged in to comp, you can ping csci). Ignore the first set of numbers. It's possible that the first number will be unusually large because of cache misses in loading the ping program. The time for cache misses doesn't truly reflect actual used network time. Run ping again for at least 10 packets. If you get large numbers this time, use them. Record the average RTT for each message size.

2. Repeat the experiment for another machine on campus but in different parts of the network (for example, ping engr or entropy), at least one machine in the U.S. but not on campus (for example, ping cs.utah.edu), and at least one machine not in the U.S (for example, www.cs.vu.nl).

   Our Internet connection goes through the Great Plains Network. If you are interested in seeing how far away we are from the machines that you pinged, you can get more information at www.greatplains.net. You can trace the route to the specific host that you pinged using traceroute on your login machine.

3. Use gnuplot, Excel, or your favorite graphing tool to graph the RTT for each message size. Draw a line for each of the four machines, all on one graph.

4. Repeat the entire set of experiments at a different time of the day.

5. Write a report on your results. Your report should be word-processed. In your report, answer the following questions:
   • Tell the IP name and location of your login machine and each machine that you pinged. For each machine that you pinged, give an estimate of the one-way latency between your login machine and the pinged machine. Justify your answer.
   • Give an estimate of the transmission rate of your login machine, and the transmission rate of the machines that you pinged. What assumptions must you make to be able to do this calculation? Justify your answer.
   • For each of the machines that you pinged, which appears to be the most significant term in RTT: latency, transmission rate, or queuing? Justify your answer.
   • Did you find ping to be a useful and reliable tool for measuring the RTT between machines? Why or why not?

Enjoy!