[Cryptography] Heartbleed and fundamental crypto programming practices

[ianG]

On 12/04/2014 19:17 pm, Tony Arcieri wrote:
> On Sat, Apr 12, 2014 at 4:15 PM, ianG <iang at iang.org
> <mailto:iang at iang.org>> wrote:
> 
>     Have you looked at QUIC?
> 
> 
> Yes (and CurveZMQ and Nitro)


:) ok...  I was just wondering if it (they) deserve mention under
Sketches, and if not why not?

http://clearcryptocode.org/tls/

And while we are on the subject, I'm curious about something:

    "Unfortunately, the modern Internet does not look favorably upon
UDP. Preliminary benchmarks of CurveCP showed nearly complete packet
loss in the face of TCP congestion which is arguably a problem with
CurveCP's decongestion algorithm but much more likely an intractable
problem with trying to do bulk transfers with UDP."


Is this over-obsessing on an edge case, aka congestion?  Most use for
most protocols is bursty therefore presumably uncongested.  As long as
this works most of the time, we've got a protocol, and the occasional
time it gets into congestion-deadly-embrace, we can handwave it off as
dead net;  happens anyway, apps deal with it, or don't.

I noticed the same long discussion in QUIC doco, and the recent post
where someone had shown CurveCP to be bad in certain tests that focused
on the congestion space;  is this a case of solving the edge cases
before proving the fundamentals?

Or is there something fundamental about the congestion issue that stops
UDP being usable under any circumstances?


iang