[Cryptography] We need a new encryption algorithm competition.

[Krisztián Pintér]

Tony Arcieri (at Tuesday, March 18, 2014, 11:50:45 PM):
> Well first, the CurveCP implementation isn't usable.

it would be rather premature to think about that. i doubt it is the
major obstacle here.

> Another part of the problem is CurveCP is built on UDP. I'm not
> really sure why it's built on UDP as djb classes TCP Vegas decongestion as "good":

it is explained here:

http://curvecp.org/availability.html

tl;dr: stream based protocols area easy to sabotage. you damage one
packet or inject a new one, and the entire connection needs to be
reestablished. by using UDP, attacker has to damage all packets to
successfully disrupt communication.

> UDP is often dropped on the floor amidst TCP congestion, as this
> experimental measurement of CurveCP demonstrates: