Um Método de Web Fingerprinting baseado em Atributos de Hardware

Abstract

Web fingerprinting is the process in which a user is, with high likelihood, uniquely identified by the extracted features from his/her device, generating a fingerprint. In order to be effective, the method must generate a stable fingerprint, and therefore it is necessary to employ discriminatory attributes with low volatility, capable of providing the same characteristics over the time. There are a variety of proposed techniques, but not all of them are capable of generating a stable fingerprint. In this work it is proposed, designed and evaluated a Web Fingerprinting method that aims to employ features that provide characteristics related to the devices’ hardware. One of the ways to achieve this objective is through the use of technologies such as HTML5 and the Web Audio API. Theses are promising technologies for Web Fingerprinting methods because they provide features related to the devices’ hardware, which reduces the extracted fingerprint’s mutability and increases the number of target devices in which the method can be executed, since HTML5 is adopted by default in the most popular web browsers. As results, it was found that the HTML5 Canvas and the Web Audio API, when employed with other attributes related to the hardware characteristics of the device, converges to a web Fingerprinting method capable of uniquely identify several users (with 90,34% of accuracy). In addition, it was found that grouping weaker attributes with more relevant ones allows the Web Fingerprinting method to extract more characteristics than to use just relevant attributes.