DIY: test de-anonymization for yourself – in the browser!
Below, you can run de-anonymization attacks and compare results on a scoreboard. Attacks run on two graphs: the grey is the background information of the attacker where user identities are known, and the blue is the anonymized network with sensitive information.
Parameter Θ controls the greediness for both algorithms: higher Θ rates will make the algorithms more precautios (less mappings with less error).
The #seeds determines the number of initial mappings before the algorithms are run.
With parameter δ you can set the sensitiveness of the Blb attack to node degrees. For example, with low δ it will not seriously consider node degree differences, but with high δ the difference between node degrees will play a significant role. We found that it is also a control of greediness, and as a rule of thumb it is good to have it around 0.1-0.5.
You can choose between three datasets to play with. The first one is quite small and an easy prey for de-anonymization (the two graphs are identical). It will surely run on older smartphones, though. Datset #2 and #3 are larger ones with varying levels of differences in the background and anonymized graphs.
to see the attacks in action – right in your browser!