[Cryptography] recommending ChaCha20 instead of RC4 (RC4 again)

[Eugen Leitl]

On Wed, Mar 12, 2014 at 09:56:54PM -0400, dan at geer.org wrote:

> Back to the conference room and the sales cycle, how many, many times
> I've had some prospective customer say something like "10% performance
> hit for security? That's a deal breaker!"
> 
> My answer: "Well, 10% is 2.5 months' worth of Moore's Law..."

Moore's law is over (no more linear semilog plots), and
performance increase has never followed Moore anyway.

A lot of software developers haven't yet realized that
multi-threading scaling is doomed, too.
 
> Sometimes it worked.  Sometimes it didn't.