Slashdot Mirror
How The FBI Might've Opened the San Bernardino Shooter's iPhone 5c (schneier.com)
"Remember the San Bernardino killer's iPhone, and how the FBI maintained that they couldn't get the encryption key without Apple providing them with a universal backdoor?" Slashdot reader LichtSpektren quotes Bruce Schneier:
Many of us computer-security experts said that they were wrong, and there were several possible techniques they could use. One of them was manually removing the flash chip from the phone, extracting the memory, and then running a brute-force attack without worrying about the phone deleting the key. The FBI said it was impossible. We all said they were wrong. Now, Sergei Skorobogatov has proved them wrong.
Sergei's new paper describes "a real world mirroring attack on the Apple iPhone 5c passcode retry counter under iOS 9." The process does not require any expensive and sophisticated equipment. All needed parts are low cost and were obtained from local electronics distributors. By using the described and successful hardware mirroring process it was possible to bypass the limit on passcode retry attempts... Although the process can be improved, it is still a successful proof-of-concept project.
Sergei's new paper describes "a real world mirroring attack on the Apple iPhone 5c passcode retry counter under iOS 9." The process does not require any expensive and sophisticated equipment. All needed parts are low cost and were obtained from local electronics distributors. By using the described and successful hardware mirroring process it was possible to bypass the limit on passcode retry attempts... Although the process can be improved, it is still a successful proof-of-concept project.
they knowed how
Is anyone REALLY surprised that the FBI was wrong? Government doesn't attract top-tier talent. Never has, never will. When your hiring practices, policies, procedures, compensation and benefits are all at the bottom of the barrel, well... that's what you get. The bottom of the barrel.
It's been common wisdom for years that with physical access to the device and unlimited time and resources, almost all encryption schemes can be defeated. In many cases this might simply mean using a mechanism to bypass the encryption rather than defeating it through brute force. But the fact is, regardless of what protections they have, devices have to ultimately present the data to the user unencrypted to actually use it. So there is usually always some kind of way in.
This attack is still done on device. It just clones the NAND back to "0 strikes" after each 6 attempts.
This attack doesn't extract the memory and doesn't decode externally. It just copies NANDs.
Why is this significant? Because it means you can't do extraction in parallel, you still have to go through all the codes sequentially on the device.
It defeats the significant portions of the backoff. It defeats the erase after n failures. It's a very significant attack.
But no one said this type of attack was impossible. I personally read about variants on this attack while the controversy was going on. I even posited it myself. I believe Apple even addressed it claiming that this attack wasn't possible on later iPhones due to a change in how the failure count is stored.
http://lkml.org/lkml/2005/8/20/95
> are essentially unbreakable, even using quantum computers
The Enigma was "unbreakable", until it was broken.
DES was unbreakable, until it was broken.
MD5 was unbreakable, until it was broken.
RSA was "unbreakable" last year. Not so much this year.
There are some new algorithms which haven't quite been completely broken just yet. Well, unless the new algorithm is used by someone who -also- allows an older algorithm, im which case the service using the new algorithm is vulnerable to DROWN.
As someone who can barely see a 0603 SMD device, I find this quite impressive. He was able to remove the flash from the board, get it to function, watch it communicate, and identify the multiple mechanisms used by the chip to communicate and where on the flash it accessed. I always suspected the way the FBI did it was a brute force attack on copies of the chip data.
Neat!
DES was unbreakable, until it was broken.
MD5 was unbreakable, until it was broken.
RSA was "unbreakable" last year. Not so much this year.
DES was actually designed to be crackable.
MD5 is not an encryption algorithm.
RSA has not been considered robustly secure for a long time, and was never considered unbreakable.
If decryption takes 1e6 times as long as encryption, the algorithm is easily crackable. If it takes 1e12 times as long, it is good enough for casual communications. 1e15 is secure against all but the most determined government sponsored crackers. If the ratio is 1e100 it is uncrackable in the life of the universe (the number of quarks in the universe is ~ 1e80). That ratio has been growing exponentially, far faster than computing hardware has been improving.
RSA was tampered with by the NSA to allow for it to be easily cracked. While we'd known there was tampering with it, the extent of that tampering wasn't known until the Snowden leaks. That said, the flaw is only with dual elliptic curve and I don't think anybody uses that anymore. Also the only thing cracked this year was RSA 220, which is 729 bits and the next you'd logically expect to see broken. My secure emails use RSA-1024 (I didn't set that up, all I do is check a checkbox that says "Secure" and the recipient needs to use their key card and PIN to decrypt it - not sure how it works for out of office emails).
Not a surprise that the US government uses RSA for secure emails but AES (designed in Belgium away from NSA tampering) for both military and confidential secret and top secret encoded data. Confidential data needs to be at least AES-128 encrypted and Secret/Top Secret AES-256 if I recall correctly. We're insulated from that stuff (our software backend handles it), all we need to know is the classification.
Brute force wouldn't work here, since after 3 tries, the phone would be locked for good. That's why the FBI first approached Apple, and later, went to a third party to get that phone unlocked. But it's interesting whether that would have been possible had Syed Farook had an iPhone 6 or something w/ a more advanced iOS, which would have precluded even that break-in
That is what I was wondering. Internally, the phone would have a combination of NAND and NOR flash (usually in MCPs). Unsoldering is next to impossible, since we are talking BGA packages, especially since in this case, the flash would have to go back into the phone for the phone to work.
But like you say, they could read out the contents of the flash w/o turning on the phone, and then run their brute force attack on the phone. Only thing - this was a 5c, and if it used his fingerprints, they'd be out of luck - unless they got a template of his fingerprints to apply on the home button. Incidentally, what if people do choose to use that to get their phones locked? What mechanisms does the FBI have? In this case, the phone user is dead, so his body was available to the FBI to extract his fingerprint, but what in the cases of other criminals who commit crimes, leave their phones behind accidentally and flee, but have protected phones due to their fingerprints being used? The FBI database may or may not have them
Of course they knew all along how to get into the phone, probably five different ways.
But all the public+media+dog had was speculation and unfortunately a big spotlight on the subject device.
Normally they work in secret and in the shadows and crack these phones all the time. But this one had everybody watching, and when everyone is watching, you do not get out your best-kept secrets and reveal them in front of the cameras. The agencies didn't want to confirm any of that by suddenly showing up with a cracked phone, thus revealing they had various techniques to do exactly what they wanted. So they tried the front door approach with Apple, and then some other approach where they can make some outside company look like the source and patsy.
Meanwhile all the much more secret techniques remain secret. Done.
But all the various bloggers and media people want to know exactly how it was done, which is exactly why they went to some effort to find a disposable way in rather than reveal their secrets. Meh. Who cares. Privacy is an illusion. If for a moment anyone thinks their iPhone is some kind of sacred secret place only they can access, well, they are fools. Nothing is secret.
Sig for hire.
This is exactly what everyone was saying at the time. The FBI didn't really give a damn about what was on the phone. All they wanted was the legal precedent for forcing companies to give up their security.
Correction: Stein says she is anti-war.
"Trump!!", the new Godwin.
http://blog.mdsec.co.uk/2015/0...
I tried to explain this to a number of people on other forums and got a surprising amount of pushback. Nice to have someone prove me right.
Is this a joke? hahahah "how the FBI opened the iPhone5c? Ohh how... after all, Apple does not help the Feds... AT ALL." Ok... CUT! Nice shot everyone... do you think they will believe it? How numb and dumb and fallen have we become... Brains, what I want you for?