Re: [core] OSCORE: Questions about Section 5.2

[Christian Amsüss <[e-mail]>]

Hello Jaro,

> > The expectation is that the shortest (zero-length) ID would be used in
> > cases wherever that's beneficial, eg. when a constrained device
> > primarily utilizes one context in which it is addressed as a server.
>
> This is an interesting optimization. I'm not too sure about the actual
> benefits though. To me this would only result in the constrained nodes being
> able to shave off a few bytes of allocation when constructing the response
> and saving their sender_id to persistent storage.

It's much more about message size. There are applications that need to
get an OSCORE message through w/o fragmentation over single link layer
messages (6TiSCH joining), and this helps fitting things in there.

> I'm implementing OSCORE on top of zephyr (not integrated into it) for an
> embedded board. Currently it isn't open source, but I asked my advisor,
> who'll forward the request to the supervisor.
> Judging from your github history I expect you ask for RIOT-OS? :)

Indeed, RIOT is what I have in mind. There have been successful cases of
code shared between Zephyr and RIOT, and I think that collaboration
could be fruitful here, if licensing permits.

At any rates, I'm looking forward to interoperate with a constrained
application in future plug tests.

Best regards
Christian

-- 
To use raw power is to make yourself infinitely vulnerable to greater powers.
  -- Bene Gesserit axiom