Websites Can Track Your Online Via Browsers

Websites Can Track Your Online Via Browsers

You might be aware of websites, banks, retailers, and advertisers tracking your online activities using different Web “fingerprinting” techniques even in incognito/private mode, but now sites can track you anywhere online — even if you switch browsers.

A team of researchers has recently developed a cross-browser fingerprinting technique — the first reliable technique to accurately track users across multiple browsers based on information like extensions, plugins, time zone and whether or not an ad blocker is installed.

Previous fingerprinting methods usually only work across a single browser, but the new method uses operating system and hardware level features and works across multiple browsers.

This new fingerprinting technique ties digital fingerprint left behind by a Firefox browser to the fingerprint from a Chrome browser or Windows Edge running on the same device.

This makes the method particularly useful to advertisers, enabling them to continue serving targeted advertisements to online users, even if they avoid them by switching browsers.

The new technique can be found in a research paper titled (Cross-)Browser Fingerprinting via OS and Hardware Level Features [PDF] by Lehigh University’s Yinzhi Cao and Song Li, and Washington University in St. Louis’ Erik Wijmans.

The cross-browser fingerprinting technique relies on “many novel OS and hardware features, especially computer graphics ones” that are slightly different for each computer.

For example, the technology can be used to identify the machine by performing 20 unique WebGL tasks while rendering 3D graphics in web browsers with carefully selected computer graphics parameters, such as texture, anti-aliasing, light, and transparency.

In total, 36 new features work independently of a particular browser, although they are not confined to one specific web browser on the machine.

The features tested currently includes time zone, number of CPU cores, GPU, hash values of GPU rendering results, plugins, fonts, audio, screen ratio and depth, WebGL, Ad blocking, canvas, cookies, encoding, and language.

The researchers provided both a practical demonstration as well as open source code online on GitHub. They performed a test which involved 3,615 fingerprints and 1,903 users and found that their method successfully identified 99.2% of users.

On the other hand, a single-browser fingerprinting technique called AmIUnique had a success rate of 90.8%.

“This approach is lightweight, but we need to find all possible fingerprintable places, such as canvas and audio context: If one place is missing, the browser can still be somehow fingerprinted. We leave it as our future work to explore the correct virtualization layer,” the paper notes.

The researchers also noted that this new cross-browser fingerprinting technique is not too bad, as in some cases, the method can be used as part of stronger multi-factor user authentications across multiple browsers.

For example, Banks can use this technique to check if a user logging into an online account is using the computer that has been used on every previous visit, making sure the login was legitimate even if the user is using a different machine to usual.

The researchers plan to present their paper at the Network and Distributed System Security Symposium scheduled for February 26 through March 1 in San Diego, California.

 

Facebooktwittergoogle_plusredditpinterestlinkedinmail
Facebooktwittergoogle_pluslinkedinrssyoutube

COMMENTS