3

Are there any known vulnerabilities in the GOST block cipher (GOST 28147-89)? Spcifically, I'm thinking about i.e. "master keys" which intentionally exist and are used i.e. by the Russian government.

I understand that although there are attacks against the algorithm itself, it is not broken. Is this a correct assumption?

However, it would not help the algorithm in terms of security if there were for example master keys which would allow the Russian government (or whomever, I'm using the Russian government as an example, since GOST is a Russian algorithm, for all I know) to access files encrypted with this algorithm.

If GOST would be used in a chain of several algorithms, i.e. AES-TWOFISH-GOST, would there be a risk of weakening AES and/or TWOFISH assuming GOST had known vulnerabilities like the one described earlier?

Also, is GOST used anywhere outside Russia or is it a strictly Russian algorithm?

Thank you very much.

1

Spcifically, I'm thinking about i.e. "master keys" which intentionally exist and are used i.e. by the Russian government.

The standard doesn't define S-boxes (substitutions tables) and this makes hard to invent such a master key, especially if you treat the S-Boxes as long term keys and keep them secret. Besides, no one researcher has demonstrated such a "master key", though there are some subspaces of key space that are weaker than the others (but the algorithm remains unbroken).

If GOST would be used in a chain of several algorithms, i.e. AES-TWOFISH-GOST, would there be a risk of weakening AES and/or TWOFISH assuming GOST had known vulnerabilities like the one described earlier?

It depends on the way you combine the algorithms into a chain, because some Meet-in-the-Middle attacks can be possible in this case. In addition, AES has 128-bit block, but GOST has 64-bit block and it will be tricky to combine them in secure way.

Also, is GOST used anywhere outside Russia or is it a strictly Russian algorithm?

Yes, besides Russia the algorithm is still widely used in Ukrain, Belarus, Kazakhstan and other ex-USSR republics. However, those countries have their own sets of GOST S-Boxes and they're moving to their own crypto standards (Kalina in Ukrain, Belt in Belarus etc.). There are some open source encryption tools like GostCrypt that are based on GOST. GOST is also implemented in OpenSSL and Crypto++ frameworks.

0

If GOST would be used in a chain of several algorithms, i.e. AES-TWOFISH-GOST, would there be a risk of weakening AES and/or TWOFISH assuming GOST had known vulnerabilities like the one described earlier?

Assuming that the keys for each of the encryption operations are secure random numbers that are chosen independently from each other, then a flaw in GOST would not weaken the security of the AES-TWOFISH portion of the pipeline.

Regarding the security of GOST itself, the cipher can be weakened by the choice of a poor replacement table. There are also known attacks that produce faster than brute-force results. As described in this answer, that is sufficient to consider the cipher broken.

Taking a step back, I question your entire approach of chaining encryption algorithms. Doing a good job in choosing and securing a key for AES will likely provide you with as much security as the triple-cipher pipeline you are proposing. Once you've encrypted data to the point of being uncrackable (eg: AES 256), adding further encryption doesn't make it more secure, only more complex. So stick with AES and a good key and you'll be more than safe enough.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.