JavaScript-based device fingerprinting mitigation using personal HTTP proxy

In this paper, we designed a mechanism to mitigate the existence of device fingerprinting process by utilizing personal HTTP proxy. For the purpose of detection and prevention of device fingerprinting, the proxy utilizes a two-stage filtering process; header-based filtering and content-based filtering. We define 6 values of attribute among with their weight based on the value of the highest entropy. In addition, we also define the ConfLevel coefficient as the weight accumulation value of the existence of each attribute. There is also defined ThresholdLevel coefficient for the purpose of determining the limit condition if there is an indication of device fingerprinting. System testing was performed by accessing Alexa's Top 125 Indonesia Website. We obtained 15.5796 as the ThresholdLevel value. Among the 125 web pages tested, it was found that 38 web pages were indicated utilizing JavaScript as the media of device fingerprinting activity.