[Cryptography] "Post-quantum encryption contender is taken out by single-core PC and 1 hour"

[Jerry Leichter]

> I used to day, “Software is like wine. It’s better when aged a bit.” Perhaps this applies to crypto algorithms, particularly those using newish math.
> 
> When starting to use these algorithms, it might be a good idea to double “encrypt” with a well-tested algorithm which is attackable with quantum attacks. For key exchange algorithms, it might be sufficient to use both algorithms in parallel (to minimize round trips) and XOR the results.
What's interesting in this case is some comments from the developers of the algorithm.  The algorithm (and much of this new post-quantum work) is based on some rather specialized mathematics.  Cryptographers are picking up on the math ... without a deep understanding of it (comment by the developer).  The attack was very straightforward and simple - to someone with deep knowledge of the mathematical literature on supersingular isogenies.  Complex stuff, and until now really the domain of specialists.

Then again, elliptic curves were just obscure to anyone not in a particular are of mathematics not so long ago.

Cryptography has moved a *long* way in the last 30 years or so - and quantum attacks are moving it even further along into deeply mathematical domains.  Cryptographers need to keep upping their game - and as cryptography comes to rely ever more on advanced mathematics, they are going to have to work closely with mathematicians with deep domain expertise.

Relevant story:  Back in the early 1970's, I was a graduate student in math.  A faculty member who one of my friends was close to left to go work at NSA.  My friend ran into him at a conference a few years later, and they talked about the NSA experience.  How was it to work in an organization where you couldn't publish or even talk about your work?  The faculty member's comment:  I can only publish internally, but I have more people carefully reading my work and taking it seriously and talking to me about it than I ever did when I was at the university.

They guy's field was algebraic geometry - exactly where such things as elliptic curves and isogenies live.  NSA has been at this for 50 years....

                                                        -- Jerry