wiki:GENIExperimenter/Tutorials/jacks/GettingStarted_PartII_ndn/Procedure/Execute/GEMINI

Version 3 (modified by tteixeir@bbn.com, 3 years ago) (diff)

--

Appendix for the Understanding the AM API using Named Data Networking Exercise

5.2 Visualizing your Named Data Networking (NDN) application using the GENI Desktop

These are instructions for the optional part of the Understanding the AM API using a Named Data Networking application tutorial. Make sure you have completed the instructions at Trying out the NDN Application before you continue.

5.2.1 Launch the GENI Desktop

  1. Find your slice in the GENI Portal and click on the GENI Desktop button

Figure 5-1 Click on the GENI Desktop button in the Slice Tools section of your Slice page.
  1. Before the GENI Desktop launches, it will ask for authorization. Follow the instructions to click on the button.

Figure 5-2 Click on the Authorize the GENI Desktop button.
  1. It will ask for you to choose your member authority.

Figure 5-3 Click on your member authority.
  1. A pop-up will ask you to Add Global Node. Go ahead and add a Global Node.
  1. The next screen will welcome you to the GENI Desktop. Wait until the tool is ready. This may take several minutes to complete. It should look like the image on the right when it is done.

Figure 5-4 GENI Desktop Welcome.

5.2.2 Setting up the visualization

We are now ready to graph traffic on the four links of the topology.

  1. Click on the Passive Graphs Lite in the menu that pops up.
  1. On the GEMINI Graphs menu that pops up, deselect all boxes and select linkoctets. We will be graphing the number of bytes (a.k.a octets) on the links.
  1. Select all of the links in the experiment: (1) Press the SHIFT button and then click on all of the links in turn to select them and (2) Click on Add Graphs with Title. For each link you will get a graph showing the number of octets going in and out of each interface on the link. (See figure.)
  2. You can close the GEMINI Graphs menu.

5.2.3 Visualizing your Experiment

  1. At this point you will see two graphs showing traffic on each of the links on your experiment. The traffic you see on the links is broadcast traffic from the switch on the rack to which these nodes connect. It is not related to your experiment.
  1. Go back to the ssh window you used to log into the PI or Experimenter nodes when you did the NDN exercise.
  2. Use the consumer.py program to fetch a namespace you had not previously fetched. This is to make sure the data will be fetched from the custodian and not from the cache at some node.
  1. Your graphs should now show traffic on the lan-1 and lan-2 links (see example below). The left part of the graphs is the network being quiet, while the spikes on the right are the Interest requests and Data packets. The first spike on lan-2 is a request from the Internet-Router node, thus no traffic on lan-1. The same request is made by one of the consumers PI or Experimenter, thus traffic only flows on lan-1. The second spike on both graphs is one of the consumers fetching data from the custodian node. Please take note that the graphs are not scaled for direct comparison and take careful note of whether the max Y values are in bytes or kilobytes.

http://groups.geni.net/geni/attachment/wiki/GENIExperimenter/Tutorials/Graphics/TrafficGraphs.png

Cleaning up

Go back to the NDN exercise to free up the resources you used in this experiment.