Slashdot Mirror
GSM Decryption Published
Hugh Pickens writes "The NY Times reports that German encryption expert Karsten Nohl says that he has deciphered and published the 21-year-old GSM algorithm, the secret code used to encrypt most of the world's digital mobile phone calls, in what he called an attempt to expose weaknesses in the security system used by about 3.5 billion of the 4.3 billion wireless connections across the globe. Others have cracked the A5/1 encryption technology used in GSM before, but their results have remained secret. 'This shows that existing GSM security is inadequate,' Nohl told about 600 people attending the Chaos Communication Congress. 'We are trying to push operators to adopt better security measures for mobile phone calls.' The GSM Association, the industry group based in London that devised the algorithm and represents wireless operators, called Mr. Nohl's efforts illegal and said they overstated the security threat to wireless calls. 'This is theoretically possible but practically unlikely,' says Claire Cranton, a GSM spokeswoman, noting that no one else had broken the code since its adoption. 'What he is doing would be illegal in Britain and the United States. To do this while supposedly being concerned about privacy is beyond me.' Simon Bransfield-Garth, the chief executive of Cellcrypt, says Nohl's efforts could put sophisticated mobile interception technology — limited to governments and intelligence agencies — within the reach of any reasonable well-funded criminal organization. 'This will reduce the time to break a GSM call from weeks to hours,' Bransfield-Garth says. 'We expect as this further develops it will be reduced to minutes.'"
The weaknesses of this algorithm are well-known and a new version that fixes those issues has been available for a long time. Now, does anyone knows whether this new version has been deployed everywhere? Who is still relying on the older version?
BTW, the algorithm used by 3G networks is different. It is based on AES and the design is publically available.
Nobox: Only simple products.
The NY Times article is missing quite a lot detail. Slashdot users might appreciate the raw video from the talk (torrent): part 1, 2, 3.
That's a strawman. You're using "obscurity" with two subtly different meanings. The OP's point is that the secret of a system should not depend on the algorithm; that is, a restatement of Kerckhoff's principle, which says that a system's security should reside in the key. When someone invokes the phrase "security through obscurity", what we mean is a system that violates Kerckhoff's principle and places essential details in the cryptosystem itself, which is far more difficult to keep secret than a key.
"Obscurity" of the key and "obscurity" of the cryptosystem are distinct concepts that shouldn't be conflated, but you did just that. Perhaps it is you who should refrain from commenting on security.
As another poster mentioned, the government can already get a wiretap easily enough without having to break the cipher.
I am sick and tired of conspiracy theories. Remember the sage advice to never attribute to malice what can be adequately explained by incompetence.
When someone who understands cryptographic security says "security through obscurity isn't security at all," they typically mean that knowledge of the algorithm shouldn't provide any significant benefit to an attacker. In other words, the exchange should be computationally secure even if attackers know the mechanism of encryption/decryption. In cases of public/private key encryption, the exchange should be computationally secure even if attackers know the public key.
The "obscurity" of a private key, for instance, isn't the obscurity that we're talking about. You either don't know that, or you're just out to rag on me (didn't get what you wanted for chanuquanchristmasolstice?). Whatever. My initial point, that A5/1 is naturally insecure (subject to known-plaintext attacks and hit by relatively-easily-generated rainbow tables) and this project highlights that, still stands.
I have no need to get into a credentials-off with someone on Slashdot, but I'll happily discuss the more technical aspects of cryptography with anyone interested/interesting, yourself included.
Honestly, I suspect that a few things are in play here:
- A5/1 is relatively easy to implement in limited hardware.
- Much of the existing infrastructure hardware has code that either sits in ASICs (this seems unlikely at this point) or bolted-into-a-box firmware that would require costly re-flashing.
- Companies aren't forced by consumers to provide genuine security.
- Most phone calls are *really* boring, and most of us honestly have nothing that we feel is worth hiding (I'm not saying that this sentiment is a good one in general).
I would like to think that the public will eventually get wise and call, globally, for the use of cryptographic algorithms that are more genuinely secure, even against government intrusion, but I know that this is next to impossible. Phone companies did a cost/benefit analysis on this one long ago and decided that the encryption that they were using was sufficient. With public awareness, the costs/benefits of modernization have changed (fractionally). In general, this is good news.
Actually GSM requires a stream cipher while your suggestion, AES, is a block cipher and would not work in this case.
Of course it is trivial to turn a block cipher like AES into a stream cipher, but its performance cannot compete with a "native" stream cipher.
You can streamcipher with AES. http://en.wikipedia.org/wiki/Block_cipher_modes_of_operation
Religion is what happens when nature strikes and groupthink goes wrong.