4 ms·

How do we know they don't? But at least with Microsoft, there is a way to remove the encryption key from them. Few years ago, I helped a person unlock their q

• schoen 11y ago
  > How do we know they don't? Maybe someone can resolve this question by asking Google, examining Android source code, finding relevant online documentation, or even monitoring what a new Android device does on the network when setting up FDE. I don't feel like this should be a big insoluble mystery for the ages; it's a basic security property of a largely open-source OS that's one of the most popular in the world. We could even ask Micah Lee to do a follow-up article directly addressing other operating systems' FDE key management behavior.
  • beeboop 11y ago
    You can't audit the firmware of the baseband processor, which is capable of sending the key to anyone, so there's not much point of auditing Android/asking Google. You cannot rely on any phone with a closed firmware baseband processor to be 100% secure. It might also come as a surprise to many than most every post-2008 Intel processor is similarly insecure: http://libreboot.org/faq/#intel http://libreboot.org/faq/#intel
    • simoncion 11y ago
      > You cannot rely on any phone with a closed firmware baseband processor to be 100% secure. That's not the question being asked. The question being asked is: "How do you know that the key for full disk encryption on Android is sent to Google?" Every technical person understands that any part of a system that has unrestricted read access to RAM can be used to thwart any FDE scheme. For the purposes of this conversation, it's not useful to talk about this. Doing so is like saying "One can be instantly vaporized by a nuclear bomb, therefore the question is irrelevant." in a conversation about whether aspirin or ibuprofen is safer for occasional relief of minor pain.
      • beeboop 11y ago
        My post does not answer that question and it wasn't trying to. The answer to that question is "We don't know". But suggesting that we do an audit of the code to find out doesn't seem productive, which is what I was responding to. That's a terrible analogy. It would be more accurate to compare this to aspirin and ibuprofen both sharing an unknown compound known to be capable of easily killing someone when the manufacturer wants it to, then suggesting we audit both aspirin and ibuprofen of their known compounds to make sure they don't have anything capable of killing someone. It doesn't really matter if it does, because you don't want to use aspirin or ibuprofen if you want to guarantee you won't die through someone's arbitrary decision, in the same way you wouldn't use Android on a phone if you want to guarantee security of your data.
        • schoen 11y ago
          Microsoft isn't trying to hide the behavior of Windows in this respect, independent of whether Windows could contain a backdoor in some part of the code. That's why The Intercept was able to report on what Windows does, by testing it, seeing what the UI says, asking Microsoft for comment, or whatever. The same methods could be used to report about Android's behavior in this respect, in exactly the same way. After all, the article isn't claiming that Windows has a backdoor or that the Pro version secretly also escrows user keys with Microsoft while claiming not to. Backdoors in devices and operating systems are an important risk and an appropriate and interesting topic for researchers and journalists to dig into, but this article is about, essentially, documented or disclosed OS behavior!