Preprocessing traces

In order to analyze the traces, they must be first preprocessed to get statistics files (pickled by Python).
Notice that you have the ability to convert those statistics files into CSV files.
To do so, you will need our analysis scripts and their dependencies.

Once the requirements are installed, assuming that your folder containing traces is at the root of the analysis scripts directory named input, the preprocessing can be launched with

./analyze.py -i input -p '' -M

Notice that this operation can take time depending on the size of the traces.
To speed up the preprocessing, you can launch x threads by adding the option -j x.

This done, you will obtain in the folder stats one pickled file per trace file containing information about Multipath TCP connections inside each file.
Some explanation about its format is described here.
This done, you can use these statistics files to generate graphs with provided scripts.

Is there any other format available?

In the case where you don't want to use the specified format and work with CSVs, you can launch this line after having preprocessed the files:

./convert_to_csv.py -s stats -c csvs

where stats is the directory where statistics files generated by analyze.py.
This will produce CSV files containing all data in the folder csvs.