[noise] noisecat: the noise swiss army knife

[Trevor Perrin]

On Wed, Feb 28, 2018 at 9:23 AM, Trevor Perrin <trevp at trevp.net> wrote:
> On Tue, Feb 27, 2018 at 4:40 PM, Gerardo Di Giacomo <gdg at fb.com> wrote:
>> Hello everyone,
>>
>> I wrote a netcat-like tool for noise protocol: noisecat - https://github.com/gedigi/noisecat
>
> Hi Gerardo,
>
> Nice!  I was going to steer you towards NoiseSocket, but looks like
> you don't need any length/framing bytes, since both sides are
> configured with knowledge of the handshake and then just forward
> streams.  So this ends up really simple...

Sorry, my comment above was nonsense, looks like you have a 2-byte
framing layer.  Of course some framing is necessary if you have
variable-length transport messages...

I guess I'd still point you towards NoiseSocket or even one of the
constructions we've considered on top of it, like [1].

These are works-in-progress but I'd like to get people experimenting
and trying to interop / standardize on these higher-level conventions.

The benefits for your case perhaps aren't much, since you may not care
about any sort of negotiation.  But using NoiseSocket framing wouldn't
cost much either, just an extra 2 bytes for handshake messages.

Trevor

[1] https://moderncrypto.org/mail-archive/noise/2018/001473.html