FROST stands for “fingerprinting remotely using OPFS-based SSD timing” and is a no-interaction attack that websites can run to access data stored on a computer. It leverages SSD activity triggered when visiting sites, since browsers use SSD-backed storage for temporary files. The attack creates a several-gigabyte file that blocks the SSD from moving temporary web data, while the website probes timing of incoming data from other sites. The collected timing data is analyzed with a machine learning model to predict additional online activity. Researchers report an 88.95% accuracy rate for predicting which sites a user would access.
"Researchers write that by using this technique, their machine learning model was able to predict which sites a user would access with an accuracy rate of 88.95 percent"
Read at Futurism
Unable to calculate read time
Collection
[
|
...
]