The results are designed to be reproduced on a machine running Ubuntu 14.04. It is recommended to run the simulations on a Google Compute Engine instance to ensure maximum consistency and reproducibility. Below are the instructions to reproduce:
This method requires that you have a Google Cloud account and associated billing account (or free trial) set up.
-
Navigate to your Google Cloud Console and click "Compute Engine > Images" on the left hand side.
-
Search for the ubuntu-1404 image and select it. Click "CREATE INSTANCE".
-
Choose a name for your instance (e.g. dctcp-ns2). Choose your desired zone to host your instance. Choose 4 vCPUs as the machine type. Make sure to check the box to allow HTTP traffic. Click "Create".
-
Install git and make:
$ sudo apt-get -y install git make -
Clone the repository:
$ git clone https://github.com/sibanez12/dctcp-ns2.git -
Install the dependencies (this will take about 8 minutes):
$ cd dctcp-ns2 && make -
Reproduce the results (this will take about 5 minutes):
$ ./run.sh
The plots will be saved in the plots/ directory. When the simulations are complete and the plots have been produced, an HTTP server will be started up automatically (may need to enter sudo password).
- Please identify the External IP address of the machine you are running the simulations on and navigate to http://<IP_ADDRESS> or to http://localhost if you are running on a local machine. Click on a link to view the corresponding figure.
You can use the run_sim.py script to create figures individually, but you will need to source the settings.sh script first.
Run, $ ./run_sim.py --help for usage information.