Here we will basically reproduce the same steps followed to deploy the BitTorrent experiment in VCT-C but now using the real Community-Lab testbed.
The basic difference is that now we will not have to create nodes but instead we will have to select the nodes that we want to use to run our experiment. Also, we will not have to configure each one of these nodes manually, but we will use a file containing the experimental data for slivers that we had previously validated in VCT-C.
After login into the Community-Lab Testbed Management, the basic steps that we have to follow before running the BitTorrent experiment in the Community-Lab testbed are the following:
It is important to notice that each of these steps could take some time because nodes poll periodically the controller and apply changes in the next successful connection, and therefore, changes on nodes do not appear immediately in the Community-Lab Testbed Management interface.
Following, we will show an example of how to perform the previous two tasks in more detail. In our example we will only use the experimental data from the Opentracker software when creating slivers, but we will have to repeat this procedure using the corresponding data to create other slivers containing different software (for example, those nodes running a Transmission daemon).
First of all, we have to go to the
SLICES –> Slices tab and click the add slice button to create a new slice. Then, we have to complete the Name and Description fields of the form.
Next, select the template that will be used by the slivers of this slice and open the file containing the sliver experimental data previously created when performing the initial test on VCT-C. In our case this file is called “bt_experiment-data.tar.gz”.
After saving the slice we are ready to create slivers by clicking the add slivers button.
Now, a list of all the available nodes within Community-Lab will be displayed. This list shows detailed information about the nodes: id number, architecture, group which the node belongs to, number of slivers running, available disk space, current state, etc. It is extremely important to check this information carefully before selecting the nodes so that we ensure that these nones meet the basic requirements of our experiment.
As you can see in the following screenshot, we selected 5 different nodes in Production state belonging to UPC group, where no other slivers are running that also have free disk space available.
After selecting the nodes we receive a confirmation message and we can now create the slivers by clicking yes, create slivers.
If we go to the
SLICES –> Slivers tab we will now see the 5 recently created slivers with names “Real_Bit-Experiment@UPC-lab104-VMxx”. In this stage, we would have to repeat this process to create other slivers containing different software. To do so, we would have to select the appropriate experimental data file (instead of the “bt_experiment-data.tar.gz” file in this example).
The next step is to set the slice state to START and wait till all the slivers are running.
As we can observe, currently there are 6 slices deployed on Community-Lab and our slice “Real_Bit-Experiment” is properly running.
Now we are ready to run our experiment and access its data.