2) If you have jurisdiction over someone, you can force them to decrypt something based on a warrant, so it isn't even desireable for a backdoor to exist. Rather than hoping they follow the rules in their ultra-secret room that normal people will never see, they can work through the court system, publicly.
3) There's a concept called "attack surface." A backdoor increases that, by increasing the number of secrets.
4) The article goes into detail over the real-world failures of systems like these.
5) The device manufacturer does not necessarily have our interests at heart and should not be the party contesting warrants in any reasonable system.
6) Even if the encryption scheme was sound, human error trumps all. We're introducing all sorts of elements, each of which is a point of failure, and none of which has interests aligned with those of the end user (us). OTP is provably secure, you know. It's the key management that gets you.
So we haven't demonstrated the need for this vs. compelling someone via warrant to decrypt and other means. We have good reason to believe that illegal, dragnet surveillance is the real reason for opposition. Even if you believe that our own government is (and always will be) perfectly virtuous with that power, there are many foreign governments with similar capabilities. Not sending them any traffic? Think again (BGP hijacks). And we have lots of experience showing that such systems leave people vulnerable to bad guys.
So this asks us to take on big risks for bad reasons and we know that. There just aren't any good reasons not to dismiss and oppose such systems outright and there are many good reasons to be suspicious of the motives of anyone telling us otherwise.
I blacklisted WaPo some time ago as not being a credible news source. This editorial only gives me further reason to keep them blacklisted.
A lot of people in IT seem to be blind to this. Encryption is no different from a very good hiding spot. If the authorities know it exists and you know how to access the information within it, there are ways to get you to disclose that information. It's not like they will just let you walk away if you simply say "no".
And that's just for the normal "criminals" argument. For "terrorists" there are even more ways to get them to disclose information. Rubber-hose cryptanalysis is a thing.
Rather than critiquing your specific solution, allow me to explain why no solution is possible in general.
0. Because this is Apple, people are assuming Apple's model, but it is not acceptable to enshrine that model into legislation for all of the reasons I could go into if they don't seem sufficiently obvious to anyone. And if Cyanogen on a Nexus device or Debian on a PC doesn't support the backdoor (or does but the device owner has the power to remove it), it is clearly unreasonable to hold Apple to a different standard, and in any event would be ineffective because the "criminals" (or anyone desiring privacy) could just use non-Apple devices.
1. Even if you are willing to pass a law against general purpose computing, if you have a backdoor that can't be updated you have both a security vulnerability (because flaws can't be fixed) and an ineffective backdoor (because if someone e.g. loses their keys there is nothing you can do about it).
2. If you have a backdoor that can be updated, whatever process is used for updating the backdoor is now a security vulnerability. Having the updates signed by the Director of the FBI and the Pope provides no security because the Director of the FBI and the Pope have neither the expertise nor the time to understand what it is they're signing, and a rubber stamp enforced by digital signatures is still a rubber stamp.
Also, what happens once a key is leaked? Are we supposed to throw away every phone and every computer?
Also, if you allow firmware to reflash its keys, then you have a mechanism that again can be subverted (as well as what you point out). I think we've seen it with games that that sort of DRM doesn't work in the long run -- it's routinely cracked, but manufacturers don't care too much as long as it allows them time-windows long enough to make bank. Making it the basis of all data-handling in the land doesn't seem particularly smart.
Saying "it's not possible to build a secure backdoor" is a cop-out.
The truth is, we can build systems so that your data can be decrypted with either your key, or a master key. And we can build systems where the master key requires multiple coordinated parties.
Such systems can be used to implement secure backdoors, but only if we can trust those parties with the master keys to do their job properly.
That requires:
- Proper key security so their keys don't get exposed
- Proper access controls so that the keys are only used for the purpose for which they're intended.
The problem is that we know with complete confidence that all of the parties involved in that process will fail at one or both of those hurdles.
This isn't a technical problem. "Secure golden keys" are technically feasible, and not even particularly hard. It's simply an issue that there is no one that you ought to trust with such a key.
Your alternative is basically saying "it's theoretically possible, but in practice impossible". If so, the theory is incomplete: it fails to account for the human factor.
We've seen law enforcement abuse every form of power they've been handed (see National Security Letters for one example), so I don't trust them to Do What's Right. Ever.
I think that as N grows, the probability that q >= 1 approaches 1. It strikes me as a self-evident feature of human nature.
[source] http://www.sciencedirect.com/science/article/pii/S1048984314...
Russian brides I suppose, in the context of this discussion?
Unlike me stealing your data, there are very legitimate reasons for law enforcement to have access to devices to carry out investigations. The warrant requirement is a check on abuse. If there's an issue regarding abuse of power (e.g.: bribing employees, intimidating judges, overly broad laws), it would be better for everyone to work to address the real problems than taking away means for law enforcement to investigate actual criminals.
Correction: it could work, in theory, if US companies were willing to sacrifice massive amounts of revenue for no reward.
The making of phones is going to get commoditized - everyone could make a phone in a few years, from parts and using open source software. Then we will have choices and never have to use spyware again.
The gov't will be able to get the encrypted input for the device mfr, and the output decrypted from that. Given this data, they will have a huge headstart in backing out a solution to what the mfr's private key is.
Even if this weren't so, it makes every major vendor a very cost-effective target for the government. They'll devote resources to either breaking their private key (as I described above), or just using straightforward human engineering techniques. This apparent extra layer of security will not last long in practice.
EDIT - in my 2nd para above, it assumes that a mfr has a single private key, which is probably not the case. But generating a unique key pair for every discrete device is both an expense problem and introduces difficulties with managing the set of keys, which makes its security even more difficult.
1) Have the device manufacturer, FBI and Federal Court system all create their own public/private key pairs. FBI and the courts publish their public keys; the device manufacturer includes their public key, the FBI's and the court's on each device.
2) Each of these organizations stores their private key in a their own central secure facility. The key is only to exist in airgapped systems, further encrypted using a secure multi-party scheme so that multiple people are needed any time the private keys are accessed. Include personnel vetting, logging, video cameras, secure vaults, etc. to add further security.
3) When the user selects their password, the storage media is encrypted using that password, and the password is stored in a separate location on the device encrypted using the FBI's public key, the device manufacturer's public key and finally the courts' public key.
4) When law enforcement needs to decrypt the phone, they request a warrant from the court. When the warrant is approved, they send a copy of the encrypted password to the court's secure facility. The court decrypts it using their private key and hands it back.
5) The FBI (or the court, whichever) brings a copy of the warrant and the partially decrypted password to the device manufacturer. If the device manufacturer wishes to contest the warrant, they go through the normal appeals process. If not, they decrypt again using their private key and hands the resulting encrypting password over to the FBI.
6) The FBI takes the resulting encrypted password back to their own secure facility and decrypts it once more using their own private key, giving them the plaintext password. They now have access to the device, and could only have done so with a valid warrant and further given a third party trusted by the suspect (the device manufacturer) a further say in whether or not to appeal the warrant.
There's my quick solution. Feel free to further the discussion by critiquing or improving it rather than downvoting to oblivion.