This is the GitHub repo for the code used in the CODASPY 2017 paper. You can obtain our (i) Netflix fingerprint database and our (ii) test capture from:
This will allow you to run the experiment that we conducted in Section 5.3.1 of the paper.
Notes:
- The Netflix fingerprint database was created in March-May 2016.
- Our traffic capture was created with adudump. For access to adudump, please contact Dr. Jeff Terrell at info@altometrics.com. If you have the source for adudump, feel free to contact us if you have issues compiling it or using it.
- Throughout this repo, we will assume that you have saved the provided files to your home directory (e.g. ~/db and ~/traffic_capture).
The easiest way to get started is to run the program on our traffic capture:
- Open a terminal, navigate to the server subdirectory, and start the server with the provided script. Make sure to specify a fingerprint dataset, like so:
./runServer.bash ~/db/codaspy_full.txt - On our machine, it takes ~20 minutes to load the dataset. Please wait until you see the message
Server startedbefore moving to the next step. Note that the server will require ~30GB of memory to run. - Open a second terminal, navigate to the client subdirectory, and run the client on a capture file, like so:
cat ~/traffic_capture/2016_09_02_traffic_capture.txt | grep ADU | awk '$6 > 200000' | cut -d" " -f2,3,4,5,6 | python preprocessor.py | python detector.py 127.0.0.1 10007 2> debug.txt 1> stdout.txt
Further information on the server and client can be obtained in their respective READMEs.