All about Browser Fingerprinting, the privacy nightmare that keeps web developers awake at night

Last week, researchers published a paper titled Browser Fingerprinting: A survey, that gives a detailed insight into what browser fingerprinting is and how it is being used in the research field and the industry. The paper further discusses the current state of browser fingerprinting and the challenges surrounding it.

What is browser fingerprinting?

Browser fingerprinting refers to the technique of collecting various device-specific information through a web browser to build a device fingerprint for better identification. The device-specific information may include details like your operating system, active plugins, timezone, language, screen resolution, and various other active settings.

This information can be collected through a simple script running inside a browser. A server can also collect a wide variety of information from public interfaces and HTTP headers. This is a completely stateless technique as it does not require storing any collected information inside the browser.

The following table shows an example of a browser fingerprint:

Source: arXiv.org

The history of browser fingerprinting

Back in 2009, Jonathan Mayer, who works as an Assistant Professor in the Computer Science Department at Princeton University, investigated if the differences in browsing environments can be exploited to deanonymize users. In his experiment, he collected the content of the navigator, screen, navigator.plugins, and navigator.mimeTypes objects of browsers. The results drawn from his experiment showed that from a total of 1328 clients, 1278 (96.23%) could be uniquely identified.

Following this experiment, in 2010, Peter Eckersley from the Electronic Frontier Foundation (EFF) performed the Panopticlick experiment in which he investigated the real-world effectiveness of browser fingerprinting. For this experiment, he collected 470,161 fingerprints in the span of two weeks. This huge amount of data was collected from HTTP headers, JavaScript, and plugins like Flash or Java. He concluded that browser fingerprinting can be used to uniquely identify 83.6% of the device fingerprints he collected. This percentage shot up to 94.2% if users had enabled Flash or Java as these plugins provided additional device information. This is the study that proved that individuals can really be identified through these details and the term “browser fingerprinting was coined”.

Applications of Browser fingerprinting

As is the case with any technology, browser fingerprinting can be used for both negative and positive applications. By collecting the browser fingerprints, one can track users without their consent or attack their device by identifying a vulnerability. Since these tracking scripts are silent and executed in the background users will have no clue that they are being tracked.

Talking about the positive applications, with browser fingerprinting, users can be warned beforehand if their device is out of date by recommending specific updates. This technique can be used to fight against online fraud by verifying the actual content of a fingerprint. “As there are many dependencies between collected attributes, it is possible to check if a fingerprint has been tampered with or if it matches the device it is supposedly belonging to,” reads the paper. It can also be used for web authentication by verifying if the device is genuine or not.

Preventing unwanted tracking by Browser fingerprinting

• By modifying the content of fingerprints: To prevent third-parties from identifying individuals through fingerprints, we can send random or pre-defined values instead of the real ones. As third-parties rely on fingerprint stability to link fingerprints to a single device, these unstable fingerprints will make it difficult for them to identify devices on the web.
• Switching browsers: A device fingerprint is mainly composed of browser-specific information. So, users can use two different browsers, which will result in two different device fingerprints. This will make it difficult for a third-party to track the browsing pattern of a user.
• Presenting the same fingerprint for all users: If all the devices on the web present the same fingerprint, there will no advantage of tracking the devices. This is the approach that the Tor Browser uses, which is known as the Tor Browser Bundle (TBB).
• Reducing the surface of browser APIs: Another defense mechanism is decreasing the surface of browser APIs and reducing the quantity of information a tracking script can collect. This can be done by disabling plugins so that there are no additional fingerprinting vectors like Flash or Silverlight to leak extra device information.

Read the full paper, to know more in detail.

DuckDuckGo proposes “Do-Not-Track Act of 2019” to require sites to respect DNT browser setting

Mozilla Firefox will soon support ‘letterboxing’, an anti-fingerprinting technique of the Tor Browser

Mozilla engineer shares the implications of rewriting browser internals in Rust