http://courses.cms.caltech.edu/cs143/

Requires python to be Python 2.7.

Installing on Ubuntu:

sudo apt-get install pip
sudo pip install -r requirements.txt

Running on Ubuntu:

bash test_all.sh
# Examine case*.pdf to see graphs

The latest network performance graphs (#96) are:

• case0_FAST_test_output.pdf
• case0_RENO_test_output.pdf
• case1_FAST_test_output.pdf
• case1_RENO_test_output.pdf
• case2_FAST_test_output.pdf
• case2_RENO_test_output.pdf

Most pull requests after #50 have performance graphs attached as comments.

Example usage:

python run.py cases/case0.json

Or, if you want to be fancy:

time python -m cProfile -s time run.py cases/case0.json case0

Run some tests with bash test_all.sh.

Usage is python run.py case_name.json [output_name]. If output_name is not provided, then it defaults to "output". Two files are generated: output_name.log and output_name.pdf. output_name.pdf has graphs of all the stats.

cases/generate_without_dot.sh generates the JSON test case descriptions. You can also generate DOT graphs to help visualize the test cases and check for errors in their descriptions. To do this:

sudo apt-get install graphviz
cd cases
bash generate.sh

Conventions

If a module involves any division, use from __future__ import division

Internal unit conventions: All time units are in seconds. All size units are in bits. So, rate units are bits per second.

Unit conventions for quantities that users specify should be consistent with these internal conventions. So, input 64 * 10**3 * 8 for a 64 KB buffer size.

Statistical information given to the user should specify its own units and doesn't need to follow the above conventions.