How's everybody doing around here? I wanted to let you know, thanks to Tony, Ron, Cees and Matthew's feedback I've gone back to the drawing board to improve my crypto knowledge.
After having battering-rammed my brain through most of Alfred J. Menezes, Paul C. van Oorschot and Scott A. Vanstone "Handbook of Applied Cryptography", I then picked up Niels Ferguson and Bruce Schneier's "Practical Cryptography" last week and have practically inhaled the first half of it in one breath. So easy and refreshing.
Most of the books and papers I have read to this point are from the ivory tower, mostly oblivious to real-world practical security issues, especially that of human comprehension and error. Worse, even after working through some of these difficult papers to get one gold "implementation nugget" I then find other material that contradicts it! For example, the envelope composition issue (MAC-then-encrypt vs. encrypt-then-MAC)..
So what's one to do, just give up? That's not an option for me, I have to move forward. I spoke with a couple of security experts at C5 and they agree with Schneier, "Cryptography is hard" and "no one can know everything about it." Therefore, at some point, I have to choose to trust some information source and go with it. I've decided to make it this 2003 book because:
1) everyone, including those on this list, seem to acknowledge Schneier as an expert 2) the book is written (as it directly claims to be) for the purpose of implementing secure crypto systems with focus on real issues. 3) seems to, more than any other source I've come across, acknowledge real-world implementation issues regarding crypto; including factoring human-frailty into the security equation (i.e., problems such as complexity). I like and agree with this philosophy.
This book (purportedly) gives the average crytologist-wannabe the advice necessary to implement secure protocols.
One idea of the book is to throw away mathematical interactions between the crypto primitives that permit certain kinds of attacks. Just a few interactions between primitives, assuming you're aware of them at all, quickly explode into many permutations very hard-to-analyze, hard-to-remember, and essentially insecure because of the hideous complexity. They therefore describe how to implement "ideal" primitives that do not suffer from these weaknesses. These implementations are typically slower than their non-ideal counterparts, but the authors claim the idea is to put security first because "there are enough fast, insecure systems out there.."
So far, I really like this book and its philosophies. Has anyone else read the book?
Cheers, Chris