Debunking a Bogus Encryption Statement?

deviantphil asks: "Recently, a coworker tried to assert that encrypting a file twice with a 64 bit algorithm is equivalent to encrypting it once with a 128 bit algorithm. I know enough about encryption to know that isn't true, but I am having difficulties explaining why and how. Doesn't each pass of the encryption create a separate file header which makes this assertion untrue? Can anyone point me to references that would better help me explain this?" What other laughable claims have you heard attributed to encryption, and how were you able to properly lay them to rest?

Comments

Snake oil FAQ

[Chuck Chunder]

Is a good place to start if you want to protect yourself against people hawking dodgy crypto.

I used to work at a small company where a guy came in and was trying to get us involved with some dodgy "crypto" product. Management were quite taken with it (he could talk and even had a patent!) but technically it was a load of horse shit with some pretty multicolour diagrams. That FAQ helped me crystalise my objections and the guy was given the heave ho before the company did anything embarrassing.

Re:Think about it as number of possibilities

[tkw954]

If you have 64 bits, that is 1.84467441 × 10^19 (2^64), meaning maximum that many tries to break the first layer of encryption. The second layer is the same number, meaning to break it would mean a maximum of 3.68934881 × 10^19 attempts.

This assumes that you can determine when you break the first layer of encryption, i.e. it won't work if the encrypted string is not distinguishable from noise. If this is not true, you must try each possible 64 bit second key for each 64 bit first key, for a maximum of 2^64*2^64=2^128 different keys, which would be equivalent to brute-forcing one 128 bit key.

DES design issues

[billstewart]

The general opinion about why NSA pushed DES to be 56 bits instead of 128 bits is that "differential cryptography" attacks weaken it to about 55 bits anyway, so in fact you're not losing anything, and the 56-bit version was more compact and easier to implement in hardware. Searching a 56-bit keyspace isn't exactly in the reach of run-of-the-mill computers - you need a whole bunch of them working together to get any speed. On the other hand, Gilmore's custom DES cracker and the distributed crack are *so* 1998. I don't know how much ASIC technology has improved since then - Pentium IIs were up to 400 MHz, compared to ~3 GHz for a typical Intel desktop today, and memory prices and performance have also improved significantly, so maybe you could use 1/10 as many machines.

Easy - but wrong...

[billstewart]

I hope you were suggesting the "Each bit doubles the strength" as one of the bogus assertions, not one of the true ones. For some kinds of algorithms, against some kinds of attacks, it's true. For algorithms like RSA and Diffie-Hellman that have some special properties to the keys, doubling the strength may require adding LogN bits. Some algorithms don't have variable-sized keys, and some of those that do aren't very good at using them - they're as strong as they are, and piling stuff on doesn't change the weaknesses, like rot-26. Some algorithms are groups - combining R rounds of N-bit keys just gets you the equivalent of one round with a different N-bit key.