Recent Progress in Cryptographic Hashing

I went to the colloquium featuring John Black and “Recent Progress in Cryptographic Hashing.” John covered the general principles behind hashing – taking a varying length message and producing an output digest of the same, fixed size. Because of this two similar messages should produce two very different hashes.

He also talked about attacks on hash functions, where you try to find two distinct inputs that will produce the same output. In particular he highlighted the work of Wang whose attack on MD5 in February of ’04 broke it. The trick was to vary 6 bit positions – now all attacks that break MD5 are based on the differences in those same bits.

John’s work has focused on improving such attacks. Where the initial break of MD5 took an hour on a supercomputer he was able to improve the attack so that it only took five minutes on a commodity computer.

Though I find security interesting, I thought that the best thing I learned from John’s talk was that it’s not only about what you are presenting (though that is a crucial part) it is also about how you present it. John did an excellent job using stories to go along with the material in his talk so it was not only informative, but also interesting.

Last modified 10 December 2007 at 7:14 pm by paul.marshall