Hi, given the new abstractions to support PolarSSL, what
interest/resistance would there be to supporting libsodium?
https://github.com/jedisct1/libsodium
For reference, the NaCL library comes from Daniel Bernstein, and is a
reference implementation of some crypto primitives in a kind of high
level library format. The crypto support is deliberately constrained
(elliptic curves, poly1305, salsa20), but the library is in turn quite
high level which it is hoped reduces the possibility for coding and
implementation errors. All of the choices are currently sensible and
sane and the algorithms are peer reviewed
Libsodium is an attempt to take the NaCL library and make it generally
available across a wide range of architectures.
Of particular interest is the use of curve25519 as public key algorithm,
it has extremely short signatures/keys, no known patent issues, decent
speed and of course elliptic curves have shown greater resistance to
cryptanalysis over the last few decades.
The motivation would be an even simpler and easier to peer review crypto
library than PolarSSL (plus it's fast and easy to code with, and I
particularly like the very short public keys/sigs)
Thanks for your thoughts
Ed W
