Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

hannesweissteiner.com

Post reply on HN
Print view View original

11–17 of 17 posts

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#11

Post by e12e »

nine_k wrote: post #5

I still have trouble understanding what information can be leaked this way. Apparently it allows to check whether a particular website was visited recently, but the article is vague in this regard. Can anybody ELI55 this?

The Ars Technica submission might be better (I've not read TFA):

https://news.ycombinator.com/item?id=48309492

They basically trained a neural network on the data they got from the SSD - and recorded data with other websites open in different tabs or even different browsers.

They could then guess/detect other open sites.

I presume, if they'd trained/recorded - they might detect other software as well.

But right now, they demonstrated (on MacOS) that if you open the exploit in a browser - they can look at SSD activity and tell you have website x, y and z open.

Might let you target users of a certain bank, child porn, regular porn, shopping sites... Mostly imagination that sets the limit.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#12

Post by SoftTalker »

Wonding about running a background program that just performs a low level of random reads and writes to the SSD, or driver-level mitigations to add random delays to disk activity, to obfuscate the contention patterns. Though I think that adding random noise to channels like this doesn't really prevent the attack, though might make it more expensive.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#13

Post by Aurornis »

SoftTalker wrote: post #12

Wonding about running a background program that just performs a low level of random reads and writes to the SSD, or driver-level mitigations to add random delays to disk activity, to obfuscate the contention patterns. Though I think that adding random noise to channels like this doesn't really prevent the attack, though might make it more expensive.

Most of us probably have some form of that going on already with the different apps and services we have running.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#14

Post by Avamander »

ttctciyf wrote: post #10

If browsers have enough low-level access to my storage hardware to carry out timing attacks for fingerprinting, it seems likely they also have enough to maliciously chug the hardware sufficiently to degrade capacity over time and otherwise impact system integrity. I hate the thought of some random website writing and overwriting random bytes in a tight loop in the background while I'm browsing elsewhere to find the c…

> [...] it seems likely they also have enough to maliciously chug the hardware sufficiently to degrade capacity over time and otherwise impact system integrity. I hate the thought of some random website writing and overwriting random bytes in a tight loop in the background while I'm browsing elsewhere to find the cause of my slow disk subsystem.

Absolutely. Things like IndexedDB get fsynced super frequently. There's no way to tell Chrome that some web apps do not need to make it do the physical disk this often.

Re: FROST: Fingerprinting Remotely using OPFS-based SSD Timing [pdf]

#16

Post by itopaloglu83 »

SoftTalker wrote: post #12

Wonding about running a background program that just performs a low level of random reads and writes to the SSD, or driver-level mitigations to add random delays to disk activity, to obfuscate the contention patterns. Though I think that adding random noise to channels like this doesn't really prevent the attack, though might make it more expensive.

Tracking people individually and selling that data is so profitable that even with the hardened security the financial incentive is significant, so maybe we should also add fines and other forms of costs to the other side of the equation as well.
Post reply on HN