The running environment for this part is the docker client, deployed as described in the handout README. We mainly developped in this one (as one of us is using a M1 chip computer). We assume the subscriptions, keys and credentials have been generated as explained in the sample run of the handout's README.
We recommand having a terminal for the the server and one for the client. Both should have PATH inside the part3 folder. You can move this folder inside the secrestroll folder of the code skeleton to be able to reach it from the client and server docker containers.
We assume running in the server container
python3 server.py runWhile in the client container, we run:
cd trace_data_extraction; chmod +x capture.sh; ./capture.shThis launches the collection of trace data with the client code for grid cell ids from 1 to 100.
The collected trace for each query is indexed by the month, day and hour of collection and stored in the folder trace_data_extraction/traces.
Bonus: the trace_data_extraction/tor_time.sh script is run in the same conditions and produces text files storing the time took for each request to be performed with and without -t option on the client to activate or not tor. This has been used to test the correlation of execution time betweeen identical queries. Note that this requires the bc package installed with apt-get install bc.
Those parts are performed by the trace_data_extraction/trace_data_extraction.ipynb file with jupyter notebook.
Note that the user will need to install the required python modules with pip install -r requirements.txt
Running all cells of this Notebook produces the features.csv file containing the feature the classifier can use.
python3 fingerprinting.pyPrints in the console the results of classification after 10-folds cross-validation using the features in features.csv file.