This repository contains materials to reproduce the results of a paper, appearing at ACM SIGCOMM 2014, which explores the "learnability" of congestion control protocols through the use of an automated protocol design tool, Remy.

You will need a C++11 compiler, git, protobuf-2.4.1, Python 2.7, and numpy.

1. Get the base ns-2 tarball and untar it.

   tar zxvf ns-allinone-2.35.tar.gz 
   

2. cd into the untarred folder.

   cd ns-allinone-2.35 
   

3. Replace the original version of ns-2.35 with our modified ns-2.35 repository. (We leave the remaining components as is.)

   rm -rf ns-2.35 
   git clone https://github.com/pratiksha/ns2.35 ns-2.35
   

4. For each figure in the paper, this repository contains a folder with a README file, which provides the SHA-1 hash for a specific commit in ns-2.35. Check out the ns-2.35 commit for the figure you wish to reproduce using the appropriate SHA-1 hash:

   cd ns-2.35
   git checkout <SHA-1-hash>
   cd ..
   

5. Now install ns-2.35 with all its components.

   ./install 
   

6. cd to the decomposability folder in order to run the simulations.

   cd ns-2.35/tcl/ex/decomposability/
   

7. Follow the instructions in the README file of the appropriate figure's folder to run the simulations and summarize the results. The graphmaker and ellipsemaker scripts in ns-2.35/tcl/ex/graphing-scripts can be used to generate the ellipse figures from the results.