Member-only story
FalkHash (Exotic Probably Shitty Hash) Makes Me Smile
Well, some of the cybersecurity research papers might make your head spin, but underneath there’s something that is perhaps a simple method that can be explained easily. Obviously, it is important to have scientific rigour, formal proofs, extensive evaluation, and so on. But, in crypto, what’s wrong with saying spin round 10 times, move things around, and shuffle things up a bit, or taking existing methods and scrambling them together, and make them fast. If there’s a test around to check the goodness of the method you have employed, then test is against that, until it passes. It could thus be argued that it was the test that was at fault if we found that things were not quite designed as they should.
In cybersecurity, we need hashing methods to provide the integrity of data. And for this, we turn to SHA-1, SHA-256 and SHA-3. These have been designed so that they do not have too many security weaknesses (and where an adversary could modify data to produce a collision to two different data sets. But, in many applications, such as in hash tables and Bloom filters, we just need a way to match some data to an index value, and where the hash can be the index. As it’s a look-up table approach, the hashing method does not have to have the same security levels, and where it obviously needs to be fast, as we do not way a delay in looking up data. And, so…
