3 ms·
Every time an article like this hits HN, one of the standard responses is how trivial this would be to check doing network analysis. I'm curious if anyone has
by lunchladydoris 9y ago
Every time an article like this hits HN, one of the standard responses is how trivial this would be to check doing network analysis.
I'm curious if anyone has ever actually gone and done the so-called trivial network analysis to check it out?
- jvehent 9y agoI don't think it would be that easy. If this is done "the right way", then facebook processes audio locally and only uploads keywords every once in a while. The traffic would be indistinguishable from normal https traffic between your phone and their servers.
- edf13 9y agoAlso - Facebook is likely doing SSL Pining so it isn't easy to unpack their SSL traffic either... Also - The Facebook App itself is heavily obfuscated making the task of rooting through the source code very difficult to try and discover what is going on... Not an easy task at all... (Although I doubt they are actually recording sound)
- rurban 9y agoNope. Their traffic and energy consumption is very distinguishable. It's like 3x higher energy usage with the Messenger App running. Don't know the exact data volume usage, but it's also significantly higher. Decrypting the SSL traffic is a bit too hard. That's why every privacy concerned citizen uses chrome with the mbasic.facebook.com url.
- pdkl95 9y agoNote that this will be hard to generalize without some type of proof that different devices are running the same software. (different automatic updates pushed to different people? Some sort of A/B test? We cannot be sure without verifying hashes of the software)